CN105574051B

CN105574051B - Update the method and processing system of the rule that user is met

Info

Publication number: CN105574051B
Application number: CN201410621107.8A
Authority: CN
Inventors: 蒲若昂; 李俊良; 彭志成; 强琦; 黄益聪
Original assignee: Alibaba Group Holding Ltd
Current assignee: Alibaba China Network Technology Co Ltd
Priority date: 2014-11-06
Filing date: 2014-11-06
Publication date: 2019-11-22
Anticipated expiration: 2034-11-06
Also published as: CN105574051A; HK1224035A1; HK1223984A1

Abstract

The application provides the method and real-time update processing system for updating the rule that user is met, pass through the data activity degree and content according to user tag, user tag is divided into dynamic user tag, user tag and the user tag of sliding of position etc. do not have to the label of classification, and different update processing methods is used for different classes of user tag, especially update dynamic user tag the most frequent is used and quickly and effectively updates processing method, improve processing capacity and processing speed that system updates request to user, and then it improves user and meets real-time update and the inquiry of rule to improve the efficiency of user data update.

Description

Update the method and processing system of the rule that user is met

Technical field

This application involves communication and computer field more particularly to one kind can be for the above user of ten million rank and opposite In the rule condition of million ranks, support to calculate the user tag of crowd's rule of user's satisfaction in real time more in second grade/Millisecond The user of new method and the application user tag update method meets rules process method and user meets rule process side Method system.

Background technique

It is widely applied in current internet by masses, people already are familiar with and are happy to consume by internet, amusement, send out Cloth personal information etc., consumption, amusement and the information of the Internet user that businessman provides by searching for service provider on the internet Record and change information, can more fully understand the demand of audient or the consumer group, it is more accurate find audient or The consumer group simultaneously launches advertisement to them or provides service.Such as data platform division department, Alibaba also provides orientation and launches Platform, wherein confined goods are applied to the wireless primary client of Taobao, are substantially carried out card and recommend to recommend with crowd.Taobao commonly sells Family can be oriented marketing in Taobao shop and loose-leaf by determining operation product, for example, Taobao details page and try out Center, to improve the profit of businessman.

As Internet user's quantity sharply increases, data platform need handle update quantity sharply increase, such as Ah In mother's data management platform (DMP platform) be crowd's precision marketing big data application platform, end 2014 it is all Mother's Ali advertising business has accessed mother's Ali data management platform, service at present thousands of advertisers, daily handle it is up to ten thousand Crowd, average daily income 5,000,000.It is expected that end of the year advertiser, which reaches 100,000, crowd's number, reaches 500,000, advertisement daily takes in 30,000,000, In addition ten million or even more than one hundred million orders of magnitude are had reached for the number of users of Taobao website, and daily user's more amount of new data also reaches Arrive more than one hundred million orders of magnitude, and the order of magnitude of rule is also more than million ranks.

It is existing meet for user provide there are mainly two types of the processing methods of rule query: one is offline batch processing method, A kind of incremental computations processing method.

Wherein, Fig. 1 is the outline flowchart that user meets Map in the offline batch processing method of rule in the prior art, Fig. 2 Meet the outline flowchart of Reduce in the offline batch processing method of rule for user in the prior art, combined with Figure 1 and Figure 2, institute Stating offline batch processing method is to realize that the offline batch processing method can using Map-Reduce (mapping-reduction) framing program It is calculated with handling billions of inputs and tens of thousands of crowds within a few hours.The basic step that the Map-Reduce framing program is realized Suddenly it is divided into two stages of Map (mapping) and Reduce (reduction): wherein in the Map stage, comprising the following steps:

Step S01 ': first reading crowd's rule list, step S03 ': parsing the regular expression of crowd, obtain with currently Map inputs the atom conditional expression of the relevant duplicate removal of table, compiles these using Fel (Fast Expression Language) The atom conditional expression of duplicate removal；

Step S04 ': input table data, the atomic expression after executing all compilings are read in one by one；

Step S05 ': after the completion of being executed for a data；

Atom conditional expression (Cid, the Clause that step S07 ': Map exports User ID, user meets in table Identity) list, each user correspond to multiple atomic expressions, and not each user's output is primary, and every corresponding former The record of subexpression all exports once.

In the Reduce stage, comprising the following steps:

Step S11 ': first and crowd's rule list (Gid list, Group Identification list) is read in；

Step S13 ': parsing the regular expression (Gid, Group Identification) of crowd, different from Map It is that is parsed here is the atom conditional expression list that each crowd needs each table to meet；

Step S14 ': the output for reading in Map is inputted as data, is closed to each user in the data that each Map is exported And the expression list that user meets on a plurality of record of each table is obtained after merging；

Step S15 ': merge above two steps as a result, the corresponding relationship of user and crowd's rule are obtained, for each User executes strictly all rules；

Step S17 ': judge whether user meets rule；

Step S19 ': output user meets the result of rule.

The mode of offline batch processing, can handle the calculating for comparing big data quantity described in the prior art, and handling capacity is big, but Delay is very big, generally requires tens hours or even longer time, is consequently belonging to the processing speed of day (Day) rank.

Fig. 3 is the outline flowchart for indexing batch processing method that user meets rule in the prior art, another in conjunction with Fig. 3 Kind method is to generally use foundation index in the way of index process and the mode of increment is combined to handle, comprising the following steps:

Step S21 ': the demographic data of full dose is read first；Step S23 ': being established by offline batch processing system and indexed, Index file is generated to be placed in distributed file system；Step S25 ': on-line system reads the index in distributed system, load Get up；Step S27 ': then, online query service is provided by API.

By way of indexing and combining incremental processing, it can obtain a result faster to an inquiry, but index Time still needs ten hours or so, therefore still belongs to the processing speed of hour (Hour) rank, and can not be to current The crowd of update carries out calculating output result in real time.

In conclusion existing two kinds of solutions, the crowd that cannot handle in real time calculates demand, and cannot handle people The real-time update of group, the function that it is even more impossible to support to inquire affiliated crowd from people.Therefore, how for huge number of users, The situation that resulting user largely updates and regular quantity is equally huge with quantity is updated, such as more than ten million rank In real time user and rule condition relative to million ranks calculate crowd's rule of user's satisfaction in second grade/Millisecond, i.e., in real time Judge that an audience objective meets the rule of which seller in real time, by the mode of the result of acquisition with higher efficiency be supplied to by Many targets provide preferably experience to audience objective, such as advertiser or seller are helped to reach conclusion of the business etc., become industry concern The problem of.

Summary of the invention

The application technical problems to be solved are can largely to update for huge number of users, resulting user And the situation that regular quantity is equally huge with quantity is updated, the efficiency of user data update is improved, is realized in second grade/Millisecond The real-time query for crowd's rule that user meets is calculated in real time, and the result of acquisition is supplied to audient in a manner of higher efficiency Target.

In view of this, the application provides a kind of method of rule that update user is met, wherein include:

The user data is divided into not by the user data for obtaining user to be updated according to the attribute of the user data Generic multiple user tags；

According to the classification information of the user tag, processing is updated to the user tag, to search the use of variation The rule that family label and the user tag for recalculating the variation are met.

Preferably, wherein the classification information of the user tag include dynamic, the dynamic user tag be attribute with Environmental change and changed user tag

Further, processing is updated to the user tag, further includes:

The mapping relations are established, mapping relations, clause's expression formula including regular expression to clause's expression formula to rule The then mapping relations of expression formula, dynamic user tag to the mapping relations of clause's expression formula, user and clause's expression formula The bitmap hit relationship whether hit and user hit the mapping relations of regular expression；

The user data of user to be updated is compared with the user data of original customer data base, searches the dynamic of variation The user tag of state is returned if do not found, is entered in next step if finding；

The dynamic user tag for searching the variation is closed in the mapping of the dynamic user tag to clause's expression formula Corresponding several clause's expression formulas in system；

For corresponding several clause's expression formulas, time whether user to be updated hits clause's expression formula is carried out Calculating is gone through, the hit relationship of the user to be updated and corresponding several clause's expression formulas are obtained；

The bitmap hit relationship is compared before the hit relationship is calculated with the traversal, finds the hit of variation Relationship；

Hit relationship for the variation is to become being not hit by from hit, then searches son corresponding to the hit relationship Sentence expression formula, and will be described right in the corresponding several regular expressions of mapping relations of clause's expression formula to the regular expression Several regular expressions answered are deleted from the mapping relations for the regular expression that the user to be updated hits；

Hit relationship for the variation is to become hit from being not hit by, then by the user to be updated and the hit The bitmap hit of clause's expression formula corresponding to relationship, which updates, to be uploaded in the bitmap hit relationship, and searches the hit relationship Corresponding clause's expression formula clause's expression formula to the regular expression the corresponding several regular expressions of mapping relations, Corresponding several regular expressions are increased in the mapping relations of the regular expression of user's hit to be updated.

Preferably, processing is updated to the user tag further include:

According to the dynamic user tag content, the dynamic user tag, which is divided into, can enumerate class, can not enumerate Class, which can be indexed, and can not be enumerated can not index class；

Attribute storage of array is established to the enumerable dynamic user tag, and is looked into using filtering query method It askes；Search tree storage is established to the dynamic user tag that can not be enumerated but can index, and is inquired using indexing method；To not It can enumerate and the dynamic user tag that can not be indexed divides range storage, and storage is numbered to range, and using can rope Draw query method and may filter that query method query composition.

Preferably, each dynamic user tag has the timestamp of record latest update time.

Further, processing is updated to the user tag further include:

Policy Updates request is received, by carrying out classified inquiry control to the dynamic user tag, establishes user institute The new mapping relations of the rule of satisfaction.

Preferably, the classification information of the user tag further includes position, the user tag of the position be attribute and when Between user tag relevant with spatial position.

Further, processing is updated to the user tag further include:

Spatial index number is established to all users, unique global identity of the user is established and states the space of foundation The mapping relations of index number；

The user data for transmitting user to be updated distributes local identification code to the user to be updated, establishes spatial index The mapping relations of number and local identification code；

The user data of the user to be updated is compared with original customer data base, if the user to be updated with Before be not present, then establish new spatial index number if the user to be updated is original user and check the spatial index Whether number changes, and to spatial index number, there is no the use of the position for then updating the user to be updated of variation The attribute information of family label, it is changed to spatial index number, delete the original local knowledge of the user to be updated Other code establishes new local identification code, and updates the mapping relations of local identification code and spatial index number.

Further, during the user data of the transmission user to be updated, using Distributed Message Queue mode Data are carried out to the user data and update decoupling, and are cut by data of the key signature character to the user data Point.

Preferably, establishing spatial index number to all users using hibert curve.

Further, spatial index is established to the user tag of the position further include:

Modulus is carried out again to the spatial index number of the foundation to divide.

Preferably, being updated processing to the user tag and including:

The attribute information of the user tag of the position is stored in continuous caching, the caching is with one to work as Ground identification code is the offset Compass of following table.

Preferably, being updated processing to the user tag further include:

According to the variation frequent degree of the user tag of the position of the user to be updated, dynamic setting updates threshold values, when The variation of the user tag of the position of the user to be updated is updated again when being more than the update threshold values.

Preferably, further including supporting data mistake during the update processing method of the user tag to the position Phase sets data expiration period by dynamic, deletes the data not updated yet more than data expiration period.

Preferably, further including the user tag of sliding to the classification of the user tag, the user tag of the sliding is Attribute itself it is constant but with environmental change changed user tag.

Preferably, being updated processing to the user tag further include:

According to the frequent degree of the user tag of sliding using different storage modes in case updating and inquiring, for frequent journey The user tag of high sliding is spent using matrix storage mode；The user tag of the sliding low for frequent degree is deposited using chained list Storage mode.

Preferably, the classification information of the user tag further includes context and static state, the user tag of the context It is attribute with the simultaneous user tag of condition, the user tag of the static state is the fixed user tag of attribute, using building Vertical function determining method handles the user tag of the context, using the user tag for establishing indexing means and handling the static state.

Present invention also provides a kind of systems of rule that update user is met, wherein includes:

First device, to obtain the user data of user to be updated, according to the attribute of the user data, by the use User data is divided into different classes of multiple user tags；

Second device is updated processing to the user tag according to the classification information of the user tag, to search The rule that the user tag of variation and the user tag for recalculating the variation are met.

Preferably, the classification information of the user tag includes dynamic, and the dynamic user tag is attribute with environment Change and changed user tag.

Preferably, the dynamic engine server second device includes the first engine, and first engine is to the dynamic User tag be updated processing:

Establish the mapping relations, comprising: mapping relations, the clause's expression formula to rule of regular expression to clause's expression formula Then the inverted list mapping relations of expression formula, the inverted list mapping relations of dynamic user tag to clause's expression formula, user and one Bitmap hit relationship and user that whether a clause's expression formula hits hit the inverted list mapping relations of regular expression；

Hit relationship for the variation is to become being not hit by from hit, then searches son corresponding to the hit relationship Sentence expression formula, and will be described right in the corresponding several regular expressions of mapping relations of clause's expression formula to the regular expression Several regular expressions answered are deleted from the mapping relations for the regular expression that the user of the user to be updated hits；

Hit relationship for the variation is to become hit from being not hit by, then by the user to be updated and the hit The bitmap hit of clause's expression formula corresponding to relationship, which updates, to be uploaded in the bitmap hit relationship, and searches the hit relationship Corresponding clause's expression formula clause's expression formula to the regular expression the corresponding several regular expressions of mapping relations, The mapping that corresponding several regular expressions are increased to the regular expression of user's hit of the user to be updated is closed In system.

Further, first engine is updated processing to the user tag, further includes:

Preferably, each corresponding timestamp for being stored with the record latest update time of the dynamic user tag.

Further, first engine is updated processing to the user tag, further includes:

Policy Updates request is received, classified inquiry control is carried out by the dynamic user tag to user, establishes user The new mapping relations of the rule met.

Preferably, the classification information of the user tag includes position, and the user tag of the position is attribute and time User tag relevant with spatial position.

Preferably, the second device further includes several second engines, described to handle the user tag of the position The framework of second engine includes:

First module, to receive and transmit the user data of the user to be updated；

Second module, to transmit and distribute the user data of the user to be updated；

Third module, to transmit simultaneously dispatch user inquiry request and Policy Updates request；And

4th module is updated and returns the result to the user tag to the position.

Further, the process that the 4th module is updated and returns the result to the user tag of the position, packet It includes:

To all users establish spatial index number, establish the user unique global identity and the spatial index The mapping relations of number；

Further, several first modules using Distributed Message Queue mode to the user tag of the position into Row data update decoupling, and carry out cutting by data of the key signature character to the user tag of position.

Preferably, during second engine establishes spatial index to the user tag of the position, the 4th mould Block establishes spatial index number to the user tag of the position using hibert curve.

Further, it during second engine establishes spatial index to the user tag of the position, also wraps It includes and modulus division is carried out again to the spatial index number of the foundation.

Preferably, second engine is updated processing to the user tag further include:

Preferably, also being wrapped during update processing method of second engine to the user tag of the position It includes, dynamic sets data expiration period, deletes the data not updated yet more than data expiration period.

Preferably, the classification information of the user tag further includes sliding, the user tag of the sliding be attribute and when Between user tag relevant with spatial position.

Preferably, the second device further includes several third engines, the third engine marks the user of the sliding The update processing method of label uses different storage modes according to the frequent degree of the user tag of sliding including the third engine In case updating and inquiring: the user tag of the sliding high for frequent degree uses matrix storage mode；It is low for frequent degree Sliding user tag use storage of linked list mode.

Preferably, the system for the rule that the update user is met further includes several mirror nodes, each described the One engine, each second engine and each third engine have at least one mirror nodes to provide service.

The classification information of the preferable user tag further includes context and static state, and the user tag of the context is For attribute with the simultaneous user tag of condition, the user tag of the static state is the fixed user tag of attribute.

Optionally, the system also includes 3rd device, the 3rd device updates processing institute using indexing means are established Static user tag is stated, the first device directly updates the processing context information label and returns result.

Preferably, the second device point duty and the corresponding user tag of absorbed responsible processing.

Compared with prior art, the method described herein for updating the rule that user is met passes through according to user tag Data activity degree and content, user tag is divided into the use of dynamic user tag, the user tag of position and sliding Family label etc. does not have to the label of class, and uses different update processing methods for inhomogeneous label, especially most to update It is frequent dynamic user tag using processing method is quickly and effectively updated, improves the place that system updates request to user Reason ability and processing speed, and then improve user and meet real-time update and the inquiry of rule to improve the effect of user data update Rate；

Further, the herein described update processing method to the dynamic user tag is by establishing several mappings Relationship is inquired and is calculated using mapping relations, is equivalent to from fine granularity and is cached to inquiry calculated result, therefore is big Amplitude accelerates calculating；It is calculated by related clause's expression formula to dynamic user tag, is carried out from calculating simultaneously Beta pruning, therefore calculation amount is greatly reduced, and then computational efficiency greatly improved, it further increases user and meets the real-time of rule It updates and inquires to improve the efficiency of user data update.

In addition, system of the application by establishing the rule that the update user is met, is calculated using distributed rule Framework, the calculation block of the user tag including Distributed Architecture and position to the division and user tag for establishing user tag Frame realizes the reasonable distribution and processing in real time of user tag, further increases the processing speed that user updates request.

Detailed description of the invention

By reading a detailed description of non-restrictive embodiments in the light of the attached drawings below, the application's is other Feature, objects and advantages will become more apparent upon:

Fig. 1 is the brief process that user meets Map in the offline batch processing method of rule in the prior art

Fig. 2 is the outline flowchart that user meets Reduce in the offline batch processing method of rule in the prior art；

Fig. 3 is the outline flowchart for indexing batch processing method that user meets rule in the prior art；

Fig. 4 is the real time updating method that the user according to one embodiment of the application meets rule to user tag Divide schematic diagram；

Fig. 5 updates the system framework schematic diagram for the rule that user meets described in one embodiment of the application；

Fig. 6 updates the second engine framework signal in the system for the rule that user meets described in one embodiment of the application Figure.

The same or similar appended drawing reference represents the same or similar component in attached drawing.

Specific embodiment

In a typical configuration of this application, terminal, the equipment of service network and trusted party include one or more Processor (CPU), input/output interface, network interface and memory.

Memory may include the non-volatile memory in computer-readable medium, random access memory (RAM) and/or The forms such as Nonvolatile memory, such as read-only memory (ROM) or flash memory (flash RAM).Memory is computer-readable medium Example.

Computer-readable medium includes permanent and non-permanent, removable and non-removable media can be by any method Or technology come realize information store.Information can be computer readable instructions, data structure, the module of program or other data. The example of the storage medium of computer includes, but are not limited to phase change memory (PRAM), static random access memory (SRAM), moves State random access memory (DRAM), other kinds of random access memory (RAM), read-only memory (ROM), electric erasable Programmable read only memory (EEPROM), flash memory or other memory techniques, read-only disc read only memory (CD-ROM) (CD-ROM), Digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices or Any other non-transmission medium, can be used for storage can be accessed by a computing device information.As defined in this article, computer Readable medium does not include non-temporary computer readable media (transitory media), such as the data-signal and carrier wave of modulation.

The core concept of the application is, real-time calculating be in order to mass data — that is, number of users, regular quantity with And update quantity huge data-this on the basis of, the data changed in real time in the short time are quickly calculated, are returned in real time It returns as a result, helping audience (such as advertiser or seller) to inquire rapidly meets crowd.The data wherein changed in real time Real-time update including user tag (such as location information, turnover etc.) and to the real-time creation of crowd's rule with more Newly.

In conjunction with above-mentioned core concept, a kind of offer user is provided in the embodiment of the application and meets the real-time of rule Update method and real-time update system.By the attribute according to user tag, the data including user tag storage it is specific in The active degree and content etc. of appearance and data, are divided into dynamic user tag, the user tag of position and cunning for user tag Dynamic user tag etc. does not have to the label of classification, and uses different update processing methods for different classes of label, especially It is the user tag of the position relative complex to update dynamic user tag the most frequent and content using fast and effective Update processing method, thus greatly improve user update request processing speed, and then improve user meet rule it is real-time It updates and inquires to improve the efficiency of user data update, audience is supplied in a manner of higher efficiency, realize audient group Crowd's rule that body meets in the nearest user of second grade/Millisecond real-time query.

Specifically, with reference to the accompanying drawing and specific embodiment is described in further detail the application.

A kind of real time updating method of user's satisfaction rule is provided in the one aspect of the application, comprising: is obtained to more The user data is divided into different classes of multiple users according to the attribute of the user data by the user data of new user Label；According to the classification information of the user tag, processing is updated to the user tag, to search user's mark of variation Sign and recalculate the rule that the user tag of the variation is met.

In the one side of the application, the user data of each user includes a global identity and multiple user tags, When user's proposition update petition, multiple user tags of user to be updated are obtained.Wherein, the preferable reality optimal in the application It applies in example, according to the attribute of the user data, can classify to the user tag, the user tag classification information Including dynamic user tag, the user tag of the user tag of position and sliding, the use of the user tag of context and static state Family label etc..Dynamic user tag is included at least to the classification of the user tag, wherein the dynamic user tag is Attribute changed user tag with environmental change.In the preferred embodiment,；The classification of the user tag is also wrapped It includes, the user tag of the user tag of position and sliding, the user tag of the position is attribute and time and spatial position phase The user tag of pass,；The user tag of the sliding is that attribute itself is constant but changed user marks with environmental change Label.；In other embodiments, can also include to the classification of the user tag, the use of the user tag of context and static state Family label., the user tag of the context is attribute with the simultaneous user tag of condition, the user tag of the static state For the user tag that attribute is fixed.

Fig. 4 is the division schematic diagram to user tag according to the application one aspect, is illustrated below in conjunction with Fig. 4, in detail The specific descriptions for illustrating user tag attribute different classes of in the preferred embodiment of the application one preferable one, in the application one In preferred embodiment, the user tag 100 is divided into following a few classes: dynamic user tag 101, position user's mark Label 103, the user tag 105 of sliding, the user tag 109 of the user tag of context 107 and static state, the dynamic user Label 101 be attribute with environmental change changed user tag 100, such as user purchase commodity, user subscription Information etc., the user tag 103 of the position are attribute user tag 100 relevant to time and spatial position simultaneously, such as User Activity position etc., the user tag 103 of the position are also hung with user's self space position because both linking up with the time Hook, therefore the calculation of the user tag 103 of position is different from dynamic user tag 101, the sliding label slippage User tag 105 is constant for attribute value itself, but with the changed user tag 100 of environmental change attribute, such as at any time Passage, (transaction stroke count attribute is constant in a period of time, but the number for stroke count of trading for the transaction stroke count in user's nearest a period of time Value changes at any time) etc..The user tag 107 of the context is attribute with the simultaneous user tag 100 of condition, Attribute occurs simultaneously with condition, such as user enjoys certain preferential difference with the means of payment and different, and the static mark The static user tag 109 of label is the fixed user tag 100 of attribute, indeclinable in the indeclinable or relatively long time User tag 100 can incorporate into as user tag 109 of static labels static state, such as age, gender, the name of user etc..Its In to the dynamic user tag 101, the user tag of position 103, sliding user tag 105 and sliding user mark Label 105 are required to be calculated in real time.The user tag 107 of the context and static user tag 109 are calculated real-time It is of less demanding.

It is described by taking the mode classification as an example in subsequent embodiment, certain those skilled in the art will be understood that The description for stating tag types is only for example, other according to data activity degree and content to user tag 100 or user data into Capable division and the method and its modification that use different processing modes to be updated processing or other division modes, such as wrap Include dynamic user tag 101, the user tag 105 of sliding combines tag types known to other skilled in the art Division methods etc., all within the thought range of the application.It is arrived according to above-mentioned category division for 100 pairs as specific user tag Can negotiate to determine by audience and developer in which specific classification, repeat no more.

In user tag 100, the dynamic user tag 101 is data storage capacity maximum, and data variation frequency is most High a kind of user tag 100, therefore how to realize that the real-time update of dynamic user tag 101 and real-time query become The key technical problem of the application, the update in the one aspect of the embodiment of the present application, to the dynamic user tag 101 The core concept that processing method is are as follows: by establishing several mapping relations, several dynamic user tags 101 and rule list Up to the hit relationship or corresponding relationship between formula, clause's expression formula and between the regular expression and clause's expression formula Mapping relations find the relevant clause's expression formula of the dynamic user tag 101 of variation using the mapping relations, And the corresponding regular expression of relevant clause's expression formula is found, relationship is hit according to relevant clause's expression formula Situation of change, the rule corresponding rule met to the dynamic user tag of the variation is updated.

In the preferred embodiment, the update processing method of the dynamic user tag 101 is specifically included:

Step S01: establishing several mapping relations, mapping relations, Zi Jubiao including regular expression to clause's expression formula Up to formula to the mapping relations of regular expression, dynamic user tag to the mapping relations of clause's expression formula, user and a son Bitmap hit relationship and user that whether sentence expression formula hits hit the mapping relations of regular expression；

Step S02: the user data of user to be updated is compared with the user data of original customer data base, is searched The dynamic user tag of variation is returned if do not found, is entered in next step if finding；

Step S03: the dynamic user tag of the variation is searched in the dynamic user tag to clause's expression formula Mapping relations in corresponding several clause's expression formulas；

Step S04: being directed to corresponding several clause's expression formulas, carries out whether the user to be updated hits clause's table Traversal up to formula calculates, and obtains the hit relationship of the user to be updated and corresponding several clause's expression formulas；

Step S05: the bitmap hit relationship is compared before the hit relationship is calculated with the traversal, finds change The hit relationship of change；

Step S06: the hit relationship for the variation is to become being not hit by from hit, then searches the hit relationship institute Corresponding clause's expression formula clause's expression formula to the regular expression the corresponding several regular expressions of mapping relations, and By corresponding several regular expressions from the mapping relations for the regular expression that the user of the user to be updated hits It deletes；

Step S07: the hit relationship for the variation is to become hit from being not hit by, then by the user to be updated Bitmap hit with clause's expression formula corresponding to the hit relationship, which updates, to be uploaded in the bitmap hit relationship, and searching should Mapping relations corresponding several rule of the clause's expression formula corresponding to hit relationship in clause's expression formula to the regular expression Corresponding several regular expressions are increased to the regular expression of user's hit of the user to be updated by then expression formula In mapping relations.

It is described herein update processing method support user dynamic user tag 101 real-time update, support from The user tag (Field) of user (User) arrives the real-time query of regular (Group), and each user is known by a unique identity Other code (ID) and multiple user tags 100 form, and the regular expression (Gid) that user meets includes multiple structures with (AND) Changing query statement, (SQL, Structured Query, the SQL statement in regular expression can split into one or more son Sentence expression formula (Cid), it may be assumed that gid-> { cid }+, each clause's expression formula only includes a relationship, wherein numerous regular expressions In clause's expression formula (Cid) much to be duplicate, such as { ' gender ': ' male ' }, { ' age ', ' 15~20 ' } etc., by building Vertical mapping relations saving inquiry calculation amount, wherein in the following description, mapping relations are including corresponding corresponding table and accordingly Inverted list etc..

Specifically, table one is the update processing method pseudocode schematic diagram of the dynamic user tag 101, to make to describe It is relatively sharp accurate, below with reference to table one, the update processing method of the dynamic user tag 101 is further retouched It states:

In the step S01, the mapping relations are established, the mapping relations include: regular expression to clause's table Up to the correspondence table (Gid-> Cids) of formula, the inverted list (Cid-> Gids) of clause's expression formula to regular expression, dynamic use The position whether inverted list (Field-> Cids), user and the clause's expression formula of family label 101 to clause's expression formula hit Figure hit relationship (User+Cid-> Bitmap) and user hit the inverted list (User-> Gids) of regular expression；

In the step S02, receive user (User) update request after, by the user data of user to be updated with it is original The user data of customer data base is compared, and searches the dynamic user tag 101 (Field) of one or more variations, such as It does not find, returns, if found, enter step S13；

In the step S03, for the dynamic user tag 101 of the variation, the dynamic of the variation is searched User tag 101 the dynamic user tag 101 to clause's expression formula inverted list (Field- > Cids) if in it is corresponding Dry clause's expression formula (Cid) is formed clause's expression formula set (Cid Set)；

In the step S04, to the inner corresponding several clause's expression formulas of clause's expression formula set (Cid Set) (Cid), the traversal calculating (Cid+User) whether the user to be updated hits clause's expression formula is carried out, the user and institute are obtained State the hit relationship (the hit relationship of User+Cid) of clause's expression formula in related clause's expression formula set；

In the step S04, by the step S14 obtain the user in related clause's expression formula set Corresponding dynamic user before the hit relationship of the hit relationship (the hit relationship of User+Cid) of clause's expression formula is calculated with traversal ((User+Cid-> Bitmap) carries out correlation ratio pair to the hit bitmap of label 101 and clause's expression formula, finds the hit of variation Relationship (hits relationship from True > False), hits relationship change if hit relationship became being not hit by now from former hit Labeled as T, if hit relationship becomes hitting that (hit relationship hits relationship from False-> True) now from being not hit by the past Variation is labeled as F；

In the step S06, corresponding clause's expression formula of T is labeled as the hit relationship change, to every height Sentence expression formula searches clause's expression formula to the inverted list (Cid-> Gids) of regular expression, obtains corresponding several regular expressions Formula (Gid), composition rule expression formula set (Gidset), by each of each described regular expression set (Gid set) Regular expression is deleted from the inverted list (User-> Gids) that the user hits regular expression；

In the step S07, to it is described hit relationship change be labeled as F clause, then will the user to be updated and The bitmap hit of clause's expression formula corresponding to the hit relationship update the bitmap hit relationship that is uploaded to ((User+Cid-> Bitmap in), and the falling in clause's expression formula to regular expression of clause's expression formula corresponding to the hit relationship is searched The corresponding several regular expressions (Cid-> Gids) of table are arranged, corresponding several regular expressions are increased to described to more In the inverted list (User-> Gids) of the regular expression of user's hit of new user.

In conclusion the herein described update processing method to the dynamic user tag 101 is by advising one Then expression formula splits into multiple clause's expression formulas, and establishes several mapping relations, is inquired and is calculated using mapping relations, phase When therefore significantly accelerating calculating in being cached from fine granularity to inquiry calculated result；Simultaneously by using dynamic Related clause's expression formula of family label 101 calculates, and beta pruning has been carried out from calculating, therefore calculation amount is greatly reduced, and big Width improves computational efficiency.

Those skilled in the art will be understood that the description of the above-mentioned update processing method to the dynamic user tag 101 It is only for example, other modifications carried out according to the intention of the update processing method of the dynamic user tag 101 of the application, including Pseudocode, the variation of programming language or corresponding control table name etc., within the thought range of the application.

Table one: the update processing method pseudocode schematic diagram of the dynamic user tag

In one embodiment another aspect of the application, to the update processing method of the dynamic user tag further include: It is stored, is specifically included according to dynamic 101 classifying content of user tag: the dynamic user tag 101 is divided for can Class is enumerated, can not enumerate index class and can not enumerate and can not index class；The enumerable dynamic user tag 101 is built Vertical attribute storage of array, and inquired using filtering query method；To the dynamic user tag 101 that can not be enumerated but can index Search tree storage is established, and is inquired using indexing method；To can not enumerate and the dynamic user tag 101 that can not be indexed is drawn Point range storage, and storage is numbered to range, and using can search index method with may filter that query method query composition.

Further, in the particular embodiment, for the enumerable dynamic user tag 101, pass through dictionary Numerical value is obtained, such as men and women's attribute, 0 and 1 can be set, then using 0 and 1 numerical value as dynamic user tag 101 One attribute of record, is put into attribute array, can merge storage for enumerable dynamic user tag 101, in turn Save memory.The dynamic user tag 101 that can not be enumerated but can index is searched by establishing search tree (B+tree) The key (Key) of tree is field value, and value (Value) is hit value, is formed hit list (Postinglist).Because of B+ tree branch It holds the upper limit (Lower_bound), lower limit searches (Upper_bound), and naturally supports range query (Range query), It is easy to be quick can to support that boolean queries (Boolean query) are therefore inquired.For the dynamic that can not be enumerated and can not index User tag 101, such as such as trade stroke count, can first pass through to range carry out range number, filtered by range number It does specific range filter again afterwards, to save operation cost, reduces and calculate the time.

In addition, each dynamic user tag 101 is newest with recording in one preferred embodiment of the application The timestamp of renewal time, since the sequence that dynamic user tag 101 updates may be indefinite, therefore by being each described Dynamic user tag 101 configures the timestamp of a record latest update time, can recorde each label it is newest more The new time, to prevent by old update request covering.Wherein, in the preferred embodiment, to each dynamic user tag 101 A timestamp is set, wherein timestamp can be made of an array, and can continuously be stored.

The herein described method for updating the rule that user is met, further includes the request of real-time reception Policy Updates, passes through Classified inquiry control is carried out to the dynamic user tag 101 of user, obtains the mapping table for establishing the user for meeting new rule.When When Policy Updates, according to new Rule content, using different querying method inquiry storage modes, different dynamic users is marked Label 101, the dynamic user tag 101 for obtaining the user of the new rule of satisfaction are gathered.By to different dynamic user marks Label 101, which establish different storage mode and inquiry mode, can largely save memory, and largely reduce inquiry calculation amount, in turn Lower query time, improves real-time.

Those skilled in the art will be understood that it is above-mentioned to the Policy Updates when to the processing side of dynamic user tag 101 The description of method is only for example, other according to the application to the Policy Updates when to the processing method of dynamic user tag 101 The modification carried out, including pseudocode, the variation of programming language, storage attribute value and storage field name etc., in this Shen Within thought range please.

In one preferred embodiment of the application, to the process of the update processing of the user tag in position, including Following steps:

S11: spatial index number is established to all users, and establishes the global identity and the Spatial Cable of the user Draw the mapping relations of number；

S12: transmitting the user data of user to be updated, distributes local identification code to the user to be updated, establishes space The mapping relations of index number and local identification code；

S13: the user data of the user to be updated is compared with original customer data base, if the use to be updated Family not previously exists, then establishes new spatial index number and check the space if the user to be updated is original user Whether index number changes, and to spatial index number, there is no the positions for then updating the user to be updated of variation User tag 103 attribute information, to the spatial index number it is changed, it is original to delete the user to be updated Local identification code, establish new local identification code, and update local identification code and the spatial index number mapping relations.

The update processing method of the herein described user tag 103 in position can satisfy the position of real-time update user Information (such as longitude and latitude, road name etc.), crowd's quantity of a certain position of energy real-time query.It is detailed below in conjunction with specific embodiment Describe the process of the bright herein described update processing to the user tag 103 in position in detail.

In the step S11, spatial index number and the global identity of the user are established to all users The mapping relations of (User ID) and the spatial index number can use hibert curve in the preferred embodiment Spatial index number (Key) is established to all users.To relatively stable user data, such as gender, possession, age bracket can be used Filter (Filter) filtering.Further, it is also possible to it is subdivided to each spatial index number progress modulus, according in system Central processor number can be divided into several piecemeals (Block), can increase parallel amount, such as the central processing according to system Device number can be divided into several piecemeals (Block), in each piecemeal, in order to increase parallel amount, carry out again to data Modulus divides, for example is divided into 10 block (specific number can be determined according to system processing capacity), each block 1/10 data for being responsible for a piecemeal update and inquiry.Wherein, the first order is stored in the index of the Block, can be used Tree table (Btree map) storage, Key are spatial index key, and value (Value) is that offset (Offset) pointer is directed toward hit list (Posting list), the method being capable of high search index efficiency.

In the step S12, the user data of user to be updated is transmitted, distributes local identification to the user to be updated Code (Local id) establishes the corresponding table (Key- > Local id) of spatial index number with local identification code.Preferably implementing In example, the user data for transmitting user to be updated can distribute to institute using second module 305 and the third module 304 Processing node processing is stated, also can use first module 309 using Distributed Message Queue (MetaQ) mode to institute's rheme The user tag 103 set carries out data and updates decoupling, and by key signature character (Tag) to the user tag 103 of position Data carry out cutting, and in an alternate embodiment of the invention, different key signature characters can pass through the identification to each user Code (ID) modulus is set, and different back end is responsible for handling a part of key signature character information, to be conveniently done The dilatation of back end and capacity reducing, and then improve the treatment effeciency that user updates request.

Further, the user has unique global identity code, when the user of the position of user marks Label 103 can be assigned the local identification code (local id) increased since 0 after entering one the 4th module 307, and length can Think 4 bytes (Bytes), while establishing the local identification code and mapping pass correspondingly with the global identity code It is that the hit list in spatial index table is stored using local identification code, by the data of the user tag 103 of the position Attribute is stored in continuous caching, and it is that lower target deviates Compass that the caching, which has with the local identification code,.

In the step S13, when user's more new data is sent in upstream, by the user data of the user to be updated with Original customer data base is compared, if the user to be updated not previously exists, establishes new spatial index number (Key), if the user to be updated is original user, check whether the spatial index number (Key) changes, to described Spatial index numbers (Key), and there is no the attributes of the user tag 103 of the position for then updating the user to be updated of variation Information, the attribute information of the user tag 103 of the position are stored in a continuous caching, can be by described in calling The local identification code (Local id) of caching is the attribute letter that the offset Compass of following table finds the user tag 103 of corresponding position Breath, so that the attribute information to user is updated, it is changed to spatial index number (Key) then delete it is described to Update the original local identification code (Local id) of user, establish new local identification code, and update local identification code with it is described The mapping relations of spatial index number.

In the preferred embodiment, by the position of the content of the user tag 103 of received position and original customer data base Update is compared in 103 content of user tag set, by the content of the user tag 103 of received position and original user data 103 content of user tag of the position in library further includes during update is compared, according to the user tag 103 of the position Variation frequent degree, dynamic setting update threshold values, and according to loading condition set update precision, according to update precision to position User tag 103 be updated.It is selected to update precision according to the loading condition of system, such as: under normal circumstances, it is greater than 1 The variation of rice all updates；If the flow of user data is big, 5 meters or more of variation can be greater than and updated again；If the number of user It is very big according to flow, then it can be greater than 20 meters and just update.Update described here refers to the update to spatial index (Key), The attribute that the location information of user such as latitude and longitude information can be used as the user tag 103 of position updates always, and influences to use The accuracy of user data.Furthermore the longitude and latitude of user changes, and the spatial index of user does not need to update under some cases, therefore The load pressure of system can be alleviated by setting update threshold value by the dynamic, improve the processing speed of system.

Those skilled in the art will be understood that the above-mentioned description for setting update threshold value to the dynamic is only for example, other roots The modification for updating threshold value and being carried out is set according to herein described dynamic, the variation etc. of range, update mode including updating threshold value, Within the thought range of the application.

It during the update processing method of the user tag 103 to the position, further include supporting data expired, Data expiration period is set by dynamic, deletes the data not updated yet more than data expiration period.In the particular embodiment, Such as if a user has reported primary information, but uploads (for example turn off application, network loss etc.) never again, then existing The information of reporting of user may be expired invalid over time, it is therefore desirable to remove these data.Specific real In existing, it can establish a thread and successively scan each piecemeal (Block), check whether every record is expired, if expired User data set can then be deleted, and Resource recovery.

In an alternate embodiment of the invention, to the update processing method of the user tag 105 of the sliding, including according to sliding The frequent degree of user tag 105 is using different storage modes in case updating and inquiring, comprising: the sliding high for frequent degree User tag 105 use matrix storage mode, can be when single user be very frequent to user using matrix storage mode Data can be traversed quickly using array, to support stochastic searching and sum；The user tag of the sliding low for frequent degree 105 use storage of linked list mode, and the time window of behavior does not store, and save storage, also save query process.

In addition, in an alternate embodiment of the invention, the update processing system further includes 3rd device 207, the user is marked The classification of label 100 can also include the user tag 107 and static user tag 109 of context, due to the context User tag 107, which can be directly used, establishes the i.e. renewable processing of function determining method, calculate in real time it is simple and easy, therefore can be by institute First device 204 is stated directly to handle and return result.The user tag 109 of the static state can use and establish indexing means Processing can be carried out indexing offline and be searched, and 3rd device 207 described in the user tag 109 of the static state is handled, and described the Three devices 207 are not high to real time handling requirement, can reduce system load and improve treatment effeciency.

In conclusion the data activity degree and content described herein by according to user tag 100, user is marked Label 100 are divided into the mark that user tag 105 of dynamic user tag 101, the user tag of position 103 and sliding etc. does not have to class Label, and different update processing methods is used for inhomogeneous label, especially to update dynamic user the most frequent Label 101 improves processing capacity and processing speed that system updates request to user using processing method is quickly and effectively updated Degree, and then improve user and meet real-time update and the inquiry of rule to improve the efficiency of user data update；Further, this Shen Please the update processing method to the dynamic user tag 101 by establishing several mapping relations, utilize mapping relations It is inquired and is calculated, be equivalent to from fine granularity and inquiry calculated result is cached, therefore significantly accelerate calculating；Together When calculated by related clause's expression formula to dynamic user tag 101, beta pruning has been carried out from calculating, therefore has substantially been subtracted Calculation amount is lacked, and then computational efficiency greatly improved, has further increased user and meet the real-time update of rule and inquire to mention The efficiency of high user data update.

In addition, the present invention also provides a kind of systems of rule that update user is met, wherein include:

Second device is updated processing to the user tag to the classification information according to the user tag, with It searches the user tag of variation and recalculates the rule that the user tag of the variation is met.

The update user that Fig. 5 updates the system for the rule that user meets described in one embodiment of the application expires The system framework schematic diagram of the rule of foot；As shown in figure 5, the system for updating the rule that user is met includes first device 204 and second device 205, in which:

User data of the first device 204 to obtain user to be updated, according to the attribute of the user data Difference is classified and is distributed to the user tag 100, and classification includes dynamic user tag, and the user tag of position is sliding Dynamic user tag etc.；

The second device 205 receives the user tag 100 that the first device 204 distributes, described to inhomogeneous use Family label 100 distributes to different engines, is updated to inhomogeneous user tag 100 using different update processing methods It returns the result, wherein the second device 205 includes at least several first engines 235, first engine 235 is to handle Dynamic user tag 101 is stated, wherein update processing method of first engine 235 to dynamic user tag 101 are as follows: build It stands between several dynamic user tags 101 and regular expression, clause's expression formula and the regular expression and institute The mapping relations for stating the hit relationship or corresponding relationship between clause's expression formula find the institute of variation using the mapping relations The relevant clause's expression formula of dynamic user tag 101 is stated, and finds the corresponding regular expression of relevant clause's expression formula Formula meets the user to be updated with user according to the situation of change for hitting relationship to relevant clause's expression formula Rule set be updated.

The system for updating the rule that user is met further includes that user updates request interface 201, rule query interface 203, the user updates request interface 201 and updates request to real-time reception user, and obtain user to be updated includes one The user data of global identity and multiple user tags 100；The rule query interface 203 receives real-time rule query request And export the result of the first device 204 real-time merger return.

In an alternate embodiment of the invention, the update processing system can also include 3rd device 207, to the user tag 100 classification can also include the user tag 107 and static user tag 109 of context.The user of the context marks Label 107 can calculate simple and easy in real time using the processing of function determining method is established, and the user tag 109 of the static state can be with Using indexing means processing is established, it can carry out indexing offline and search, therefore the user tag 107 of the context can be direct It is handled by first device 201, the user tag 109 of the static state can use the not high 3rd device 207 of real time handling requirement It is handled, reduce system load and improves treatment effeciency.

Those skilled in the art will be understood that the above-mentioned description to the 3rd device 207 is only for example, other are according to this The modification that the user tag 107 of context and the user tag 109 of the static state are carried out when applying to the Policy Updates, Including the variation such as title, storage address, within the thought range of the application.

In the preferred embodiment, each engine in the second device 205 point is blamed and absorbed responsible processing is corresponding User tag 100.Because of the diversification of user tag 100, user is being supported in use, the real time processing system needs in future It wants to support several hundred thousands of labels, the differentiated services of user tag 100 is supported by establishing, make user tag 100 can be with Accomplish horizontal extension, separately maintain, system has good scalability and maintainability, can also support different types of label It calculates and optimizes.Therefore, the corresponding engine of each label only handles oneself responsible label, if rule is responsible for not comprising engine Label, then the engine will not be processed；Equally, if user property does not include the user tag that engine is responsible for, with regard to the engine It does not need to be updated and calculate, accomplishes that each engine is ultimate attainment absorbed responsible yet.Meanwhile focusing on task at hand for cooperation engine, it establishes There is unified entrance to application layer, engine can choose the user tag of oneself calling, so that real time processing system be made to unite One supports, without building proprietary system for specific user.

In optimal embodiment, the second device 205 has multiple engines, and different engine types handles different User tag includes at least the first engine 235, and first engine 235 is to handle the dynamic user tag 101.In In preferred embodiment, the second device 205 can also include the second engine 215 and third engine 255, second engine 215 user tag 103 to handle the position, user tag of the third engine 255 to handle the sliding 105.The dynamic user tag 101, the user tag of position 103, the definition of the user tag 105 of sliding and its attribute are retouched State the definition and its attribute description phase of the identical user tag 100 in the real time updating method for meeting rule with the user Together, it repeats no more.

In one preferred embodiment of the application, the second device 205 will be divided into following a few classes: the first engine 235, Second engine 215, third engine 255.It is described by taking the mode classification as an example below in an example, certain this field Technical staff will be understood that the description of above-mentioned tag types is only for example, other are according to data activity degree and content to described What two devices 205 carried out divides and handles method and its modification that different user tags is updated processing, such as is divided into First engine 235, the second engine 215 combine division methods of tag types known to other skilled in the art etc., all Within the thought range of the application.As for specific several second devices 205 how allocation processing which user tag 100 can To negotiate to determine by audience and developer, repeat no more.

In user tag 100, the dynamic user tag 101 is data storage capacity maximum, and data variation frequency is most High a kind of user tag 100, therefore the reality of dynamic user tag 101 how is realized using first engine 235 Shi Gengxin and real-time query become the key technical problem of the application, in the one aspect of the embodiment of the present application, described first Engine 235 meets method described in the real time updating method of rule using herein described user to the dynamic user tag 101 are updated processing.

In conclusion herein described first engine 235 is logical to the update processing method of the dynamic user tag 101 It crosses and a regular expression is split into multiple clause's expression formulas, and establish several mapping relations, looked into using mapping relations It askes and calculates, be equivalent to from fine granularity and inquiry calculated result is cached, therefore significantly accelerate calculating；Pass through simultaneously Related clause's expression formula of dynamic user tag 101 is calculated, beta pruning has been carried out from calculating, therefore meter is greatly reduced Calculation amount, and computational efficiency greatly improved.

Fig. 6 user described in one embodiment of the application meets the rule that the real-time update system update user of rule meets The block schematic illustration of second engine in system then.It is in the preferred embodiment, described to update what user was met in conjunction with Fig. 6 The system of rule includes the second engine 215, and the framework of second engine 215 includes the first module 309, the second module 305, the Three modules 304 and the 4th module 307, wherein first module 309 is to transmit the request user user's to be updated User data；User data of several first modules 305 to transmit and distribute the request user user to be updated, institute If stating third module 304 to request to give dry-cure node to user query request and Policy Updates；And the 4th module 307 It is updated and returns the result to the user tag 103 to position.Wherein, the 4th module 307 establishes all users The mapping relations of spatial index number and the global identity and spatial index number；Transmission request user is to be updated The user data of user distributes local identification code to request user user to be updated, establishes spatial index number and locality The mapping relations of identification code；The user data of the request user user to be updated is compared with original customer data base, If request user user to be updated not previously exists, new spatial index number is established, if the request user waits for more New user is original user, then checks whether the spatial index number changes, to spatial index number there is no The attribute information of the user tag 103 of the position for then updating the request user user to be updated of variation, to the Spatial Cable It is changed to draw number, deletes the original local identification code of request user user to be updated, establishes new local identification Code, and update the mapping relations of local identification code and spatial index number.

The framework of second engine 215 can also include updating user data interface 301 and retrieval user interface 303, The user data for updating user data interface 301 to receive request user user to be updated；The retrieval user interface 303 to receive the request of retrieval user query and Policy Updates request.

In the present embodiment, first module 205, the third module 304 and the 4th module 307 all with it is external Service platform 313 is connected, such as comprehensive service operation and management platform (OSS), and to show processing status, statistical report is more New result and managing process.

Herein described second engine 215 can satisfy location information (such as the longitude and latitude, road that processing in real time updates user Road title etc.), and crowd's quantity of a certain position of energy real-time query.It is described in detail below in conjunction with specific embodiment herein described Process of second engine 215 to the update processing of the user tag 103 of position.

Update of herein described second engine 215 to the user tag 103 of position specifically establishes all users The mapping relations of spatial index number and the global identity (User ID) and spatial index number, preferable In embodiment, hibert curve can be used, spatial index number (Key) is established to all users.To relatively stable user Data, such as gender, possession, age bracket can be filtered with filter (Filter).Further, it is also possible to each spatial index Number progress modulus is subdivided, according to the central processing unit number of system, can be divided into several piecemeals (Block), Neng Goujia Big parallel amount, such as according to the central processing unit number of system, several piecemeals (Block) can be divided into, in each piecemeal, In order to increase parallel amount, modulus division is carried out to data again, for example is divided into 10 block (specific number can root Determined according to system processing capacity), 1/10 data that each block is responsible for a piecemeal update and inquiry.Wherein, described The index of Block stores the first order, and tree table (Btree map) storage can be used, and Key is spatial index key, is worth (Value) It is directed toward hit list (Posting list) for offset (Offset) pointer, the method being capable of high search index efficiency.

In the step S12, the user data of user to be updated is transmitted, distributes local identification to the user to be updated Code (Local id) establishes the mapping relations (Key- > Local id) of spatial index number with local identification code.Preferable real It applies in example, the user data for transmitting user to be updated can distribute to the 4th module 307 processing using the first module 309, Data more new explanation can also be carried out using using user tag 103 of Distributed Message Queue (MetaQ) mode to the position Coupling, and cutting is carried out by the data of key signature character (Tag) to the user tag 103 of position, in an alternate embodiment of the invention, Different key signature characters can be set by identity code (ID) modulus to each user, different back end It is responsible for a part of key signature character information of processing, to be conveniently done dilatation and the capacity reducing of back end, and then improves and use Family updates the treatment effeciency of request.

In the preferred embodiment, the system for updating the rule that user is met is additionally provided with mirror nodes, described Each computing engines have at least one mirror nodes to provide service in second device 205, serve a computing engines The data of multiple mirror nodes and service are impartial, for example, in one embodiment, each computing engines are with there are three mirror image sections Point, the data of each mirror nodes and service are impartial, so if one of mirror nodes break down, another two Mirror nodes can provide same service.Specifically, the information preservation of the mirror nodes, in ZK frame, path name is advised It then can be with are as follows: root/ partition number/mirror image number, wherein the contents value of ZK frame joint storage is a protobuf object, Include the information such as rpc ip/port.Mirror nodes need to keep ZK heartbeat, when inquiry when, according to mirror nodes information into Row inquiry, if a mirror nodes do not connect, can inquire next mirror nodes to live.

Wherein, the protocolbuf object (hereinafter referred to as PB) is a kind of format of data exchange of Google, it is independent In language, independently of platform.Google provides the realization of three kinds of language: java, c++ and python, each realization all includes The compiler and library file of corresponding language.It is faster than using xml to carry out data exchange since it is a kind of binary format It is many.It can be it for the data exchange under the data communication or isomerous environment between Distributed Application.As a kind of effect Rate and compatible all very outstanding binary data transmission format can be used for such as network transmission, configuration file, data storage Equal numerous areas.

Those skilled in the art will be understood that the above-mentioned description to the mirror nodes is only for example, other are according to the application The modification that the function of the mirror nodes is carried out, storage mode including use, using variation of frame etc., in the application Thought range within.

In conclusion system of the application by establishing the rule that the update user is met, using distributed rule Computing architecture handles the Distributed Architecture of user tag, according to the data activity degree of user tag 100 and content by user Label 100 is divided into user tag 105 of dynamic user tag 101, the user tag of position 103 and sliding etc. and does not have to class Label, establishes second device 205, and the second device 205 includes several first engines 235, the second engine 215 and third engine 255, different processing methods is used to inhomogeneous user tag, realizes the reasonable distribution and processing in real time of user tag, Further increase the processing speed that user updates request.

Meanwhile the herein described system for updating the rule that user is met proposes a kind of frame of second engine 215, Second engine 215 receives update request and the user data for updating user and retrieving user, is distributed processing to user, And use and spatial index is established to each user and is stored according to the spatial index, when needing to carry out location updating, It calls spatial index to be updated processing to the user tag of the position of user, and then improves the processing energy of the second engine 215 Power and treatment effeciency.

Obviously, those skilled in the art can carry out various modification and variations without departing from the essence of the application to the application Mind and range.In this way, if these modifications and variations of the application belong to the range of the claim of this application and its equivalent technologies Within, then the application is also intended to include these modifications and variations.

It should be noted that the application can be carried out in the assembly of software and/or software and hardware, for example, can adopt With specific integrated circuit (ASIC), general purpose computer or any other realized similar to hardware device.In one embodiment In, the software program of the application can be executed to implement the above steps or functions by processor.Similarly, the application Software program (including relevant data structure) can be stored in computer readable recording medium, for example, RAM memory, Magnetic or optical driver or floppy disc and similar devices.In addition, hardware can be used to realize in some steps or function of the application, example Such as, as the circuit cooperated with processor thereby executing each step or function.

In addition, a part of the application can be applied to computer program product, such as computer program instructions, when its quilt When computer executes, by the operation of the computer, it can call or provide according to the present processes and/or technical solution. And the program instruction of the present processes is called, it is possibly stored in fixed or moveable recording medium, and/or pass through Broadcast or the data flow in other signal-bearing mediums and transmitted, and/or be stored according to described program instruction operation In the working storage of computer equipment.Here, including a device according to one embodiment of the application, which includes using Memory in storage computer program instructions and processor for executing program instructions, wherein when the computer program refers to When enabling by processor execution, method and/or skill of the device operation based on aforementioned multiple embodiments according to the application are triggered Art scheme.

It is obvious to a person skilled in the art that the application is not limited to the details of above-mentioned exemplary embodiment, Er Qie In the case where without departing substantially from spirit herein or essential characteristic, the application can be realized in other specific forms.Therefore, no matter From the point of view of which point, the present embodiments are to be considered as illustrative and not restrictive, and scope of the present application is by appended power Benefit requires rather than above description limits, it is intended that all by what is fallen within the meaning and scope of the equivalent elements of the claims Variation is included in the application.Any reference signs in the claims should not be construed as limiting the involved claims.This Outside, it is clear that one word of " comprising " does not exclude other units or steps, and odd number is not excluded for plural number.That states in device claim is multiple Unit or device can also be implemented through software or hardware by a unit or device.The first, the second equal words are used to table Show title, and does not indicate any particular order.

Claims

1. a kind of method for updating the rule that user is met, wherein include:

The user data is divided into inhomogeneity according to the attribute of the user data by the user data for obtaining user to be updated Other multiple user tags；

According to the classification information of the user tag, processing is updated to the user tag, to search user's mark of variation Sign and recalculate the rule that the user tag of the variation is met.

2. the method for claim 1, wherein the classification information of the user tag includes dynamic, dynamic user's mark Label be attribute with environmental change changed user tag.

3. method according to claim 2, wherein be updated processing to the user tag further include:

Several mapping relations are established, mapping relations, clause's expression formula to rule list including regular expression to clause's expression formula Up to the mapping relations of formula, dynamic user tag to the mapping relations of clause's expression formula, user and clause's expression formula whether The bitmap hit relationship of hit and user hit the mapping relations of regular expression；

The user data of user to be updated is compared with the user data of original customer data base, searches the dynamic of variation User tag is returned if do not found, is entered in next step if finding；

The dynamic user tag of the variation is searched in the mapping relations of the dynamic user tag to clause's expression formula Corresponding several clause's expression formulas；

For corresponding several clause's expression formulas, the traversal the meter whether user to be updated hits clause's expression formula is carried out It calculates, obtains the hit relationship of the user to be updated and corresponding several clause's expression formulas；

The bitmap hit relationship is compared before the hit relationship is calculated with the traversal, and the hit for finding variation is closed System；

Hit relationship for the variation is to become being not hit by from hit, then searches clause's table corresponding to the hit relationship It, and will be described corresponding up to formula in the corresponding several regular expressions of mapping relations of clause's expression formula to the regular expression Several regular expressions are deleted from the mapping relations for the regular expression that the user of the user to be updated hits；

Hit relationship for the variation is to become hit from being not hit by, then by the user to be updated and the hit relationship The bitmap hit of corresponding clause's expression formula, which updates, to be uploaded in the bitmap hit relationship, and it is right to search the hit relationship institute The clause's expression formula answered clause's expression formula to the regular expression the corresponding several regular expressions of mapping relations, by institute State corresponding several regular expressions increase to the user to be updated user hit regular expression mapping relations in.

4. method as claimed in claim 3, wherein be updated processing to the user tag further include:

According to the dynamic user tag content, the dynamic user tag, which is divided into, can enumerate that class, can not enumerate can rope Class can not be indexed by drawing class and can not enumerating；

Attribute storage of array is established to the enumerable dynamic user tag, and is inquired using filtering query method；It is right It can not enumerate but the dynamic user tag that can be indexed establishes search tree storage, and be inquired using indexing method；To can not piece It lifts and the dynamic user tag that can not be indexed divides range storage, and storage is numbered to range, and look into using that can index Inquiry method and it may filter that query method query composition.

5. method as claimed in claim 4, wherein each dynamic user tag has the record latest update time Timestamp.

6. method the invention according to any one of claims 2 to 5, wherein be updated processing to the user tag further include:

Policy Updates request is received, by carrying out classified inquiry control to the dynamic user tag, user is established and is met Rule new mapping relations.

7. the method for claim 1, wherein the classification information of the user tag further includes position, the position User tag is attribute user tag relevant to time and spatial position.

8. the method for claim 7, wherein be updated processing to the user tag further include:

Spatial index number is established to all users, unique global identity of the user is established and states the spatial index of foundation The mapping relations of number；

The user data for transmitting user to be updated distributes local identification code to the user to be updated, establishes spatial index number With the mapping relations of local identification code；

The user data of the user to be updated is compared with original customer data base, if before the user to be updated not In the presence of then establishing new spatial index number, if the user to be updated is original user, check the spatial index number Whether change, to spatial index number, there is no the user of the position for then updating the user to be updated of variation marks The attribute information of label, it is changed to spatial index number, delete the original local identification of the user to be updated Code, establishes new local identification code, and updates the mapping relations of local identification code and spatial index number.

9. method according to claim 8, wherein during the user data of transmission user to be updated, using point Cloth message queue mode carries out data to the user data and updates decoupling, and by key signature character to the number of users According to data carry out cutting.

10. method according to claim 8, wherein establish spatial index number to all users using hibert curve.

11. method as claimed in claim 10, wherein establish spatial index to the user tag of the position further include:

12. the method as described in any one of claim 7 to 11, wherein being updated processing to the user tag includes:

The attribute information of the user tag of the position is stored in continuous caching, there is the caching one to be known with locality Other code is the offset Compass of following table.

13. the method as described in any one of claim 7 to 11, wherein be updated processing to the user tag and also wrap It includes:

According to the variation frequent degree of the user tag of the position of the user to be updated, dynamic setting updates threshold values, when described The variation of the user tag of the position of user to be updated is updated again when being more than the update threshold values.

14. the method as described in any one of claim 7 to 11, wherein be updated processing to the user tag and also wrap It includes:

Data expiration period is set by dynamic, deletes the data not updated yet more than data expiration period.

15. the method for claim 1, wherein the classification information of the user tag further includes sliding, the sliding User tag be attribute itself it is constant but with environmental change changed user tag.

16. method as claimed in claim 15, wherein be updated processing to the user tag further include:

According to the frequent degree of the user tag of the sliding using different storage modes in case updating and inquiring, for frequent journey The user tag that degree is higher than the sliding of preset value uses matrix storage mode；Use for frequent degree lower than the sliding of preset value Family label uses storage of linked list mode.

17. the method for claim 1, wherein the classification information of the user tag further includes context and static state, institute The user tag for stating context is attribute with the simultaneous user tag of condition, and the user tag of the static state is fixed for attribute User tag, using the user tag for establishing function determining method and handling the context, using establishing indexing means processing institute State static user tag.

18. a kind of system for updating the rule that user is met, wherein include:

First device, to obtain the user data of user to be updated, according to the attribute of the user data, by the number of users According to being divided into different classes of multiple user tags；

Second device is updated processing to the user tag according to the classification information of the user tag, to search variation User tag and recalculate the rule that the user tag of the variation is met.

19. system as claimed in claim 18, wherein the classification information of the user tag includes dynamic, dynamic user Label be attribute with environmental change changed user tag.

20. system as claimed in claim 19, wherein the second device includes the first engine, and first engine is to institute It states dynamic user tag and is updated processing, comprising:

Establish mapping relations, comprising: mapping relations, the clause's expression formula to regular expression of regular expression to clause's expression formula Inverted list mapping relations, the inverted list mapping relations of dynamic user tag to clause's expression formula, user and clause's table Whether the bitmap hit relationship and user hit up to formula hit the inverted list mapping relations of regular expression；

Hit relationship for the variation is to become being not hit by from hit, then searches clause's table corresponding to the hit relationship It, and will be described corresponding up to formula in the corresponding several regular expressions of mapping relations of clause's expression formula to the regular expression Several regular expressions are deleted from the mapping relations for the regular expression that the user to be updated hits；

Hit relationship for the variation is to become hit from being not hit by, then by the user to be updated and the hit relationship The bitmap hit of corresponding clause's expression formula, which updates, to be uploaded in the bitmap hit relationship, and it is right to search the hit relationship institute The clause's expression formula answered clause's expression formula to the regular expression the corresponding several regular expressions of mapping relations, by institute In the mapping relations for stating the regular expression that corresponding several regular expressions increase to user's hit to be updated.

21. system as claimed in claim 20, wherein first engine is updated processing to the user tag, also Include:

22. system as claimed in claim 20, wherein each dynamic user tag correspondence is stored with a record most The timestamp of new renewal time.

23. the system as described in any one of claim 20 to 22, wherein first engine carries out the user tag Update processing, further includes:

Policy Updates request is received, classified inquiry control is carried out by the dynamic user tag to user, establishing user expires The new mapping relations of the rule of foot.

24. system as claimed in claim 18, wherein the classification information of the user tag includes position, the position User tag is attribute user tag relevant to time and spatial position.

25. system as claimed in claim 24, wherein the second device further includes several second engines, to handle The framework of the user tag that rheme is set, second engine includes:

4th module is updated and returns the result to the user tag to the position.

26. system as claimed in claim 25, wherein the 4th module is updated simultaneously the user tag of the position The process returned the result, comprising:

Spatial index number is established to all users, the unique global identity and the spatial index for establishing the user are numbered Mapping relations；

27. system as claimed in claim 26, wherein several first modules are using Distributed Message Queue mode to institute The user tag that rheme is set carries out data and updates decoupling, and is carried out by data of the key signature character to the user tag of position Cutting.

28. system as claimed in claim 26, wherein establish space to the user tag of the position in second engine During index, the 4th module is established spatial index to the user tag of the position using hibert curve and is compiled Number.

29. system as claimed in claim 28, wherein establish space to the user tag of the position in second engine It further include carrying out again modulus to the spatial index number of the foundation to divide during index.

30. the system as described in any one of claim 25 to 29, wherein second engine carries out the user tag Update processing further include:

31. the system as described in any one of claim 25 to 29, wherein second engine carries out the user tag Update processing further include:

32. the system as described in any one of claim 25 to 29, wherein second engine carries out the user tag Update processing further include:

Dynamic setting data expiration period, deletes the data not updated yet more than data expiration period.

33. system as claimed in claim 18, wherein the classification information of the user tag further includes sliding, the sliding User tag be attribute user tag relevant to time and spatial position.

34. system as claimed in claim 33, wherein the second device further includes several third engines, and the third is drawn Hold up the processing of the update to the user tag, comprising:

The third engine is according to the frequent degree of the user tag of sliding using different storage modes in case updating and inquiring: right Matrix storage mode is used in the user tag that frequent degree is higher than the sliding of preset value；For frequent degree lower than preset value The user tag of sliding uses storage of linked list mode.

35. system as claimed in claim 20, wherein the system also includes several mirror nodes, each described first draws It has held up at least one mirror nodes and service is provided.

36. system as claimed in claim 25, wherein the system also includes several mirror nodes, each described second draws It has held up at least one mirror nodes and service is provided.

37. system as claimed in claim 34, wherein the system also includes several mirror nodes, each third is drawn It has held up at least one mirror nodes and service is provided.

38. system as claimed in claim 18, wherein the classification information of the user tag further includes context and static state, The user tag of the context is attribute with the simultaneous user tag of condition, and the user tag of the static state is solid for attribute Fixed user tag.

39. system as claimed in claim 38, wherein the system also includes 3rd device, the 3rd device is used and built Vertical indexing means update the user tag for handling the static state, and the first device directly updates the user for handling the context Label simultaneously returns result.

40. system as claimed in claim 18, wherein the second device is responsible for and is absorbed in be responsible for the corresponding user's mark of processing Label.