• 专利附录
  • 专利说明
  • 权利要求
  • 类似技术
  • 同族专利
  • 引用文献
  • 法律状态
  • 优先权
    • 专利摘要:
      COMPOSITE TERM INDEX FOR GRAPHIC DATAIt is an indexing system for graphic data. In private deployments, the indexing system provides replica index and denormalization functionality to improve interrogation performance.
    公开号:BR112013016926A2
    申请号:R112013016926-5
    申请日:2011-11-30
    公开日:2020-10-27
    发明作者:Sanjeev Singh；Bret Steven Taylor；Paul Buchheit；James Norris；Tudor Bosman；Benjamin Darnell
    申请人:Facebook, Inc.；
    IPC主号:
    专利说明:

    “COMPOSITE TERM INDEX FOR GRAPHIC DATA” i FIELD OF THE INVENTION The present disclosure generally refers to databases and, more particularly, to a data indexing system for graphic data structures.
    BACKGROUND OF THE INVENTION Computer users are able to access and share vast amounts of information through various local and wide area computer networks that include private networks as well as public networks such as the Internet. Typically, a web browser installed on a user's computing device facilitates access and interaction with information located on various identified network servers, for example, associated uniform resource locators (URLs). Conventional approaches that allow the sharing of user-generated content include various technologies or information sharing platforms such as social networking websites. Such web sites may include, be linked to or provide a platform for applications that allow users to view web pages created or customized by other users where visibility and interaction with such pages by other users is managed. by some characteristic set of rules. Such social network information, and most information in general, is typically stored in relational databases. In general, a relational database is a collection of relationships (often called tables). Relational databases use a set of mathematical terms, which can use Structured Interrogation Language (SQL) database terminology. For example, a relationship can be defined as a set of trios that have the same attributes. A trio usually represents an object and information about that object. A relationship is usually described as a table, which is organized into rows and columns. In general, all data referenced by an attribute are in the same domain and conform to the same restrictions. The specific relational model states that the relationship trios have no specific order and that the trios, in turn, do not impose order on the attributes. Access data by applications by specifying interrogations that use operations to identify trios, identify attributes, and combine relationships. Relationships can be modified and new trios can supply explicit values or be delivered from a question. Similarly, interrogations identify many trios for updating or deleting. Each trio of a relationship must be uniquely identifiable by some combination (one or more) of its attribute values. This - the combination is called the primary key. In a relational database, all data is stored and accessed through relationships. The relationships that store
    Relational databases, as implemented in relational database management systems, have become a predominant choice for storing information in databases used for, for example, financial records, manufacturing information and logistics, personal data and other applications. As computational power increased, the inefficiencies of relational databases, which previously made them impractical, were weighed by their ease of use for conventional applications. The three leading open source deployments are MySQL, PostgreSQL, and SQLite. MySQL is a relational database management system (RDBMS) that works with a server that provides multiple users with access to numerous databases. The "M" in the acronym for the popular LAMP software stack refers to MySQL. Its popularity for use with web applications is essentially tied to the popularity of PHP ("P" in LAMP). Several high traffic web sites use MySQL for data storage and user data logging. A database index is a data structure that improves the speed of data retrieval operations on a database table. A database index can be created using one or more columns from a database table, which provides the basis for both quick random observations and efficient access to requested records. The disk space required to store the index is typically less than that required by the 'table (since the indexes usually contain only the main fields according to which the table should be arranged, and exclude all other details in the table ), producing the possibility of storing indexes in memory for a table whose data is too large to store in memory.
    Indexes can be deployed using a variety of data structures. Popular indices include balanced trees, B + trees and dashes.
    A chart is an abstract representation of a set of objects in which at least some pairs of objects are connected by links. Interconnected objects are commonly called nodes, and links that connect nodes are called borders. Modeling data in a graphical structure, however, poses challenges for scalability and performance. The queries that require the cross-section of a graphic structure can require many database observations. Highly scalable systems typically rely on caching and indexing to improve interrogation response times and overall performance.
    SUMMARY The present invention provides methods, devices and systems aimed at an indexing system for graphic data. In private deployments, the indexing system provides replica index and denormalization functionality to improve performance.
    described in more detail below in the detailed description and in conjunction with the following figures.
    BRIEF DESCRIPTION OF THE DRAWINGS Figure 1 illustrates an example of an indexing system architecture according to an implementation of the invention. Figure 2 illustrates an example of a computer system architecture. Figure 3 provides an example of a network environment. Figure 4 shows a flowchart that illustrates an example of a method for adding a new object to a graph and composite index.
    DESCRIPTION OF EXEMPLIFICATIVE MODALITIES The invention is now described in detail with reference to some modalities of the same as illustrated in the accompanying drawings. In the following description, several specific details are presented in order to provide an in-depth understanding of the present disclosure. However, it is evident for an element versed in the technique that this disclosure can be put into practice without any or all of these specific details. In another example, the steps and / or process structures have not been described in detail in order not to unnecessarily occlude the present disclosure. Furthermore, although the disclosure is described in conjunction with the particular modalities, it should be understood that 'this description is not intended to limit the disclosure to the described modalities. On the contrary, the description is intended to cover alternatives, modifications and equivalents as may be included in the spirit and scope of the disclosure as defined by the appended claims. In particular deployments, the present invention is directed to a database index infrastructure that provides flexible search capabilities for objects — data and associations between data objects. The particular modalities refer to an indexing system for storing and serving information modeled as a graph that includes nodes and borders that define associations or relationships between the nodes that the edges connect in the graph. In particular modalities, the graph is, or includes, a social graph, and the indexing system is part of a larger networked system, infrastructure or platform that allows an integrated social network environment. In the present disclosure, the social network environment can be described in terms of a social graph that includes information from a social graph. In fact, the particular modalities of the present disclosure depend on, exploit or make use of the fact that most or all of the data stored for the social network environment can be represented as a social graph. The private modalities provide an economic infrastructure that can be efficient, intelligent and successful to nrediize the ascala enm 6 niúimaro evhenentially growing jisnarios of the environment.
    In particular modalities, the distributed indexing system and back-end infrastructure described in the present invention provide one or more of: low scale latency, a lower cost per request, an ease of use work structure for developers, an infrastructure that allows combined queries involving both associations (borders) and objects (nodes) of a social graph as described by way of example in the present invention, an infrastructure that provides a flexible and expressive interrogation model for objects stored and associations, and an infrastructure that is easy to call directly from PHP. Additionally, as used in the present invention, "or" can mean "and" as well as "or;" that is, "or" does not necessarily occlude "and, unless explicitly established or implicitly implied.
    Particular modalities can operate in a wide-area network environment, such as the Internet, which includes multiple systems accessible per network. Figure 3 illustrates an example of a network environment, in which several examples of modalities can operate. Network cloud 60 generally represents one or more interconnected networks, through which the systems and hosts described in the present invention can communicate. Network cloud 60 may include wide-area packet-based networks (such as the Internet), private networks, wireless networks, satellite networks, cellular networks, paging networks and the like. As Figure 3 illustrates, particular modes can operate in a 'network environment comprising a social network system 20 and one or more client devices 30. Client devices 30 are functionally connected to the network environment via a network service provider, wireless carrier or any other suitable means.
    In one example of a modality, the social network system 20 comprises computer systems that allow users to communicate or otherwise interact with each other and access content, such as user profiles, as described in the present invention. The social network system 20 is a network-accessible system that, in several examples of modalities, comprises one or more physical servers 22 and data storage 24. The one or more physical servers 22 are functionally connected to the computer network 60 through, for example, a set of routers and / or network switches 26. In an exemplary embodiment, the functionality hosted by one of the other physical servers 22 may include web or HTTP servers, FTP servers, as well as, without limitation, web pages and applications deployed using the Common Port Interface (CGI) script, PHP Hypertext Preprocessor (PHP), Active Server Pages (ASP), Hypertext Markup Language (HTML), Extensible Markup Language (XML), Java, JavaScript, Asynchronous JavaScript and XML (AJAX), and the like.
    The physical servers 22 can host functionality directed to operations Aa alotama Amrada aesqansinl IN Dar maia da avamnla The sasial sm rada sicta ID nrAaAdo hace.
    client 30, view and post information, as well as communicate with each other through the website. In the present invention, servers 22 may be called server 22, although server 22 may include several servers that host, for example, the social network system 20, as well as other content distribution servers, data stores , and databases. Data storage 24 can store content and data related to and that allow the system to operate on a social network as digital data objects. A data object, in particular deployments, is a digital information item typically stored or incorporated into a data file, database or record. Content objects can take many forms, including: text (for example, ASCII, SGML, HTML), images (for example, jpeg, tif and gif), graphics (based on vector or bitmap), audio, video ( for example, mpeg), or other multi-media, and combinations thereof. Content object data can also include executable code objects (for example, executable games within a browser or frame window), podcasts, etc. Logically, data storage 24 corresponds to more than a variety of separate and integrated databases, such as relational databases and object-oriented databases, which maintain information as an integrated collection of logically related records. or files stored on one or more physical systems.
    Structurally, data storage 24 can generally include one or more of a large class of data management and storage systems. In particular modes, data storage 24 can be deployed by any suitable physical system that includes components, such as one or more database servers, mass storage media, media library systems, area networks. storage, data storage clouds, and the like. In an example - modality, data storage 24 includes one or more servers, databases (for example, MySQL), and / or data stores. The data store 24 can include data associated with different users and / or client devices 30 of a social network system 20. In particular modalities, the social network system 20 maintains a user profile for each user of the system 20. The profiles user include data describing users of the social network, which may include, for example, first names (first, middle and last of a person, a trademark and / or company name of a business entity, etc.) biography, demographics and other types of descriptive information, such as work experience, educational background, hobbies or preferences, geographic location and additional descriptive data. For example, user profiles can include a user birthday, relationship status. city of residence and the like. System 20 can additionally store
    Relationship training can indicate users who have similar or common work experience, group members, hobbies, or educational background. A user profile can also include privacy settings that manage access to user information by other users.
    The client device 30 is generally a computer or computing device that includes functionality to communicate (for example, remotely) over a computer network.
    The client device 30 can be a desktop computer, laptop computer, tablet, personal digital assistant (PDA), navigation system in or out of the car, smart phone or other cell or mobile phone, or mobile game device, among other suitable computing devices. Client device 30 can run one or more client applications, such as a web browser (for example, Microsoft Windows Internet Explorer, Mozilla Firefox, Apple Safari, Google Chrome, and Opera, etc.), to access and viewing content over a computer network. In private deployments, client applications allow a client device user 30 to enter addresses of specific network resources to be retrieved, such as resources hosted by the social network system 20. These addresses can be Locators Uniform Resource, or URLs. In addition, once a page or other resource has been recovered, client applications can provide access to other pages or records when the user "clicks" hyperlinks to other resources. For example, such hyperlinks can be located within web pages and provide an automated way for the user to enter the URL of another page and retrieve that page. Figure 1 illustrates an example of the modality of a networked system, architecture or infrastructure 100 (hereinafter referred to as the present invention and networked system 100) that can implement the social network system backend functions 20 illustrated in Figure 3. In particular embodiments, networked system 100 allows users of networked system 100 to interact with each other through social networking services provided by networked system 100 as well as with third parties. For example, users on remote user computing devices (for example, personal computers, netbooks, multimedia devices, cell phones (especially smart phones), etc.) can access the networked system 100 via web browsers. or other user client applications to access web sites, web pages, or web applications hosted or accessible, at least in part, by the networked system 100 to view information, store or update information, communicate information, or otherwise interact with other users, third-party websites, web pages, or web applications, or other infarenas are armazanada LEsaenadada to anrmceílval aAr elietama am rads 41060 In meaAaliAas represent users, concepts, topics, and other information (data) , has' chart borders that connect or define relationships between chart nodes, as described in more detail below.
    Referring to Figure 1, in particular modalities, the networked system 100 includes numerous client or web servers 104 (hereinafter referred to as the present invention client servers 104) that communicate information to and from networked system users 100. For example, users on remote user computing devices can communicate with client servers 104 through load balancers or other suitable systems through any suitable combination of networks and service providers.
    Client servers 104 can interrogate the index and database systems described in the present invention in order to retrieve the data to generate structured documents to respond to user requests. Networked system 100 may also comprise an index layer comprising one or more index servers 106, a cache layer 108 comprising one or more cache servers, and a database layer comprising one or more more database servers and associated database management functionality 110. Database 110 generally connotes a system database that can properly include other layers of cache to handle other types of queries. Each of the client servers 104 7 communicates with a cache layer 108. Cache layer 108 can be deployed as one or more distributed cache rings or groupings. In a deployment, cache layer 108 is a read / write cache layer, where all reads and writes go through the cache layer. In a deployment, the cache layer maintains association information and, thus, can handle queries for such information. Other queries are passed to database 110 for execution. In particular embodiments, database 110 is a relational database.
    Database 110 can be deployed as MySQL, and / or any suitable relational database management system such as, for example, HAYSTACK, CASSANDRA, among others. In particular embodiments, layer 108 may include an operational plug-in to interoperate with any suitable database deployment 110. In a deployment, a plug-in performs multiple translation operations, such as translation of data stored in the cache layer as graph nodes and borders for queries and commands suitable for a relational database that includes one or more tables or flat files.
    In particular modalities, the information stored by the networked system 100 is stored within the database 110 and the cache layer 108. In nartinsmlarãe modes informing armaanada dantro of each hancre de dades 1410 information is stored by the cache layer in the form of a graph that includes graph nodes and associations or connections between nodes (referred to in the present invention as graph borders).
    In particular modalities, each node or graph object is assigned a unique identifier (ID) (hereinafter referred to as the ID node) that uniquely identifies the graph node in the graph; that is, each ID node is globally unique. In a deployment, each node ID is a 64-bit identifier. In a deployment, a fragment is an allocated segment of the ID node space. In particular modalities, the graph can maintain a variety of different types of nodes, such as users, pages, events, posts, comments, photographs, videos, previous information, concepts, interests and any other element that would be useful to represent as a knot. Border types correspond to associations between nodes and can include friends, followers, subscribers, fans, worshipers (or other indications of interest), posting, comments, links, suggestions, recommendations, and other types of as- —socializations between we. In a deployment, a portion of the graph can be a social graph that includes user nodes that correspond to a respective user in the 'social network' environment. The social graph can also include other nodes such as concept nodes intended or directed to a particular concept as well as topic nodes, which may or may not be ephemeral, intended or directed to a particular topic of current interest among users of the environment social network. In particular modalities, each node has, represents or is represented by a corresponding web page ("profile page") hosted or accessible in the social network environment. For example, a user node can have a corresponding user profile page on which the corresponding user can add content, make statements, and otherwise express himself. By way of example, as will be described below, several web pages hosted or accessible in the social networking environment such as, for example, user profile pages, user profile pages, or topic profile pages, allow users post content, post status updates, post messages, post comments that include comments on other posts entered by the user or other users, declare interests, declare a "like" (described below) for any of the aforementioned posts as well as specific pages and content , or otherwise express or perform various actions (hereinafter, these and other user actions may collectively be called "posts" or "user actions"). In some embodiments, the post may include a link to, or otherwise, reference to additional content, such as media content (for example, photos, videos, music, text, etc.), uniform resource locators (URLs ), and other nodes, through their respective profile pages, other user profile pages, pages
    Such posts, statements, or actions can then be viewable by the author user as well as other users. In particular modalities, the social graph additionally includes a plurality of borders that define or represent a connection between a corresponding pair of nodes in the social graph. As discussed above, each content item can be a node in the chart linked to other nodes.
    As described above, in several examples of modalities, one or more of the web pages or web applications described are associated with a social networking environment or social networking service. As used in the present invention, a "user" can be a individual (human user), an entity (for example, a company, commerce or third party), or a group (for example, of individuals or entities) that interacts or communicates with or through such a social network environment. As used in the present invention, a "registered user" refers to a user who has been officially registered within the social network environment (In general, the users and user nodes described in the present invention refer to registered users only, although it is not necessarily a requirement in other modalities, that is, in other modalities, the users and user nodes described in the present invention can refer to users who have not been registered with the social network environment described in the present invention). In particular modalities, each user has a corresponding "profile" page stored, hosted or accessible by the social network environment and viewable by all or a selected subset of other users. In general, a user has administrative rights to all or a portion of their respective profile page as well as, potentially, to other pages created by or for the particular user that includes, for example, home pages, web applications that host pages, among other possibilities. As used in the present invention, an "authenticated user" refers to a user who has been authenticated by the social network environment as the claimed user on a corresponding profile page for which the user has administrative rights or, alternatively, a representative of the claimed user's trust. A connection between two users or concepts can represent a defined relationship between users or concepts in the social network environment, and can be defined from the point of view of an appropriate data structure of the social network environment as an edge between nodes corresponding to users, concepts, events, or other nodes in the social network environment for which the association was made. As used herein, a "friendship" represents an association, as a defined social relationship, between a pair of users in the social networking environment. A "friend", as used here, may refer to - any user of the social networking environment with which another user has formed a connection, the friendly relationship will be a relationship that causes an edge to be aerated between them explicitly as, for example, by one of the two users selecting the other for friendship as a result of transmission, or causing it to be transmitted, a friend request to the other user, who can accept or deny the request.
    Alternatively, friendships or other connections can be automatically established.
    Such social friendship may be visible to other users, especially those who are friends with one or both registered users.
    A friend of a registered user may also have increased access privileges to content, especially content declared or generated by the user, on the profile of the registered user or another page.
    It should be noted, however, that two users who have a friendship connection established between themselves on the social graph may not necessarily be friends (in the conventional sense) in real life (outside the social network environment). For example, in some deployments, a user may be a company or other non-human entity, and thus
    unable to be friends with a human user in the traditional sense of the word.
    As used herein, a "fan" can refer to a user who is an advocate or follower of a particular user, web page, web application, or other web content accessible in the social networking environment.
    In particular modalities, when a user 'is a fan of a particular web page ("fans" of the particular web page), the user may be listed on that page as a fan for other registered users or the general public to see .
    Additionally, a user's avatar or profile photo may be mustard on the page (or on any of the pages described below). As used here, a “like” can refer to something, like, by way of example and not as a limiting factor, a post, a comment, an interest, a link, a fragment of media (for example, photo , photo album, video, music, etc.), a concept, an entity or a page, among other possibilities (in some deployments, a user can indicate or declare a like for, virtually, anything in any page hosted by or accessible by the social network environment or system), that a user, and particularly a registered or authenticated user, has declared or has otherwise demonstrated that he / she likes, is a fan of, defends, appreciates, or otherwise has a positive opinion.
    In a modality, indicating or declaring a "like" or indicating or declaring that the user is a "fan" of something can be processed and defined equivalently in the social network environment and can be used in a way that can be changed; similarly, declaring yourself a "fan" of something, such as a concept or concept profile page, or declaring that an individual "likes" something, can be defined equivalently in the social network environment and used in a passable way changes in this document.
    In addition, as used herein, an "interest" may refer to an interest declared by user, as an interest declared by user shown on the user's profile page.
    As used herein, a "wish" can refer to something that what a user can declare or otherwise demonstrate an interest in, "like, or have a relationship with, such as, for example, a sport, a sports team, a musical genre, a musical composer, a hobby, a business (company), an entity, a group, a celebrity, an individual who is not a registered user, or even an event, in some modalities, another user (for example, an unregistered user), etc. As an example, there may be a concept node and concept profile page for "Jerry Rice", the famous professional football player, created and managed by one or more of a plurality of users (for example, other than Jerry Rice), although the social graph additionally includes a user node and user profile page for JeremyRice created by and managed by Jerry Rice himself (or Jerry Rice's legal or authorized representatives).
    In an exemplary graphic structure, a data object includes a plurality of attributes. The attributes can be name and value pairs. For example, a data object corresponding to an individual can include the following attributes: ("id": 12345, t G4bit FBid '"type": person, f can be a name full or nickname "created": 1253665137, '"name": "Papa Smurf'," username ":" papa smurf ', "gender": "male", "emails": [' psmurf (Qfacebook.com "," papasmurfQgmail.com "]) The data object identifier (id) can be a 64-bit value that is assigned —when the object is created. The attributes of the data object can be analyzed and maintained in a Search Index maintained by a or more index servers 106. For example, when a new data object is created, a term producer module can create the following terms from the preceding data object: type: person created: i253665i37 name: papa name: smurf username: papa smurf gender: male) emails: psmurf (Q facebook.com emails: papasmurf (Qgmail.com
    (such as a 32-bit counter or clock value) and the data object identifier (id) of a corresponding data object.
    The terms can be stored in one or more Indexes in association with a corresponding document.
    For example, in an exemplary search index, docids are generated from the object ID and the "created" data and time stamp so that all posting lists are sorted in reverse chronological (conceptually, docids are "created (32bits): OBid (64bits)"). The data stamp and time stamp (created) corresponds to the time when the data object was first created.
    In other deployments, the data and time stamp can correspond to the time that a given data object was finally modified.
    In a deployment, the docids for the index are constructed so that the results of a given survey can be ordered in reverse chronological order by time of creation.
    For example, based on this scheme, the search, name: smurf type: individual, will find all individuals that have the name "smurf", chronologically ordered in reverse order by the time in which the data object associated with the individual was created.
    In other embodiments, an arbitrary —32-bit selection key can be used in place of a data and time stamp, if desired to sort objects on some other basis. 'Associations (edges) between objects can be conceptually modeled and stored as data objects - called "edge objects". Consequently, the 'Index can store entries corresponding to data objects, such as individuals, and other objects that correspond to edge relationships that facilitate search by social network or other graph-related information, thus increasing system performance.
    The following data object corresponds to a "fan" association between the individual object above (id 12345) and another data object (id 67890) corresponding to a musical group (Coldplay). (
    "id": 92821,
    "type": conection.fan,
    "created": 1253665248,
    "source": 12345, 4 Papa Smurf
    "dest": 67890 H Coldplay
    )
    Borders can generate special terms in the search index associated with the source and target objects.
    The search query conection.fan.to (67890), for example, will display the document identifiers associated with all Coldplay fans (docid
    - 67890). Similarly, the search query conection.fan.from (12345) presents
    '13/20 all status updates from friends or other connections of an individual to the query: conection.from (12345) type: status As an additional example, the following search query presents all of an individual's friends (id 12345) who are also Coldplay fans: conection.friend.from (12345) conection.fan.to (67890) Since it is possible to make a data object "point" to another object directly with an attribute, certain types of associations can be created without a separate border object. For example, instead of having "owner" border objects between situation messages and users, a situation object can include an owner attribute with a data object value corresponding to the creation user - for example: ( "id": 5834639, "type": "status.message", "text": "doing nothing",. "owner": 12345,:) 'In a deployment, index servers 106 support simple graphical syntax cross-sectional through the interrogation composition. For example, the following search query will show an individual's friends of friends (id 12345): conection.friend.from (conection.friend.from (12345)). In that case, an Index server first performs the internal query, conection.friend.from (12345). The document identifiers presented by the internal query are then applied to the external prefix so that the entire expression expands to an OR of terms co-connection.from to all of the individual's friends.
    This interrogation composition syntax can be used to interpret a wide variety of interrogations. For example, the following search will show all photos that have tags that identify friends of the individual (id 12345) and on the Stanford network: —conection.tag.from (conection.friend.from (12345) network: stanford) type: photo . The internal interrogation syntax can be applied to any properties, not just borders. For example, assuming the "author" is an attribute of status messages, the following search would show all status messages from the individual's friends (id 12345): au- thor (conection.friend.from (12345)) type: status. In addition, the following research question will present all status messages from individuals with the Pope name: au- db ml mA AMADA Bamesseasnsv tomecsatadiaos Altos disessa 4 minmeessasAaA3Za give mass «» - mAsAINDP ca reverse: source (conection.friend.from (12345 )) type: conection.
    Index server 106 presents document identifiers in response to queries, which a client process 104 can use to access corresponding data objects stored in a data store, such as database 110 or cache layer 108. In a deployment, the term producer module, as discussed above, generates terms for the search index of attributes of data objects. A term producer takes an object as an input and issues a set of (docid, term) pairs indicating which terms should be indexed for that data object. In a deployment, the type or behavior of the term producer module is chosen based on a type of the object being inserted. For example, assume, for didactic purposes, that a term producer module has been called to process the following border object: ("id": 92821, "type": "conection.fan",. " created ": 1253665248,:" source ": 12345,% Papa Smurf." dest ": 67890 t Coldplay) The connection term producer module can produce the following terms for insertion into the index: (92821," type: conection .fan ") dé Edge document (92821," source: 12345 ") (92821," dest: 67890 ") (12345," conection.fan.to:67890 ") authentic Papa Smurfs document (12345," conection.to: 67890 ") (67890," conection.fan.from: 12345 ")% * Coldplay's document (67890," conection.from: 12345 ") Figure 4 shows an exemplifying method associated with new objects of creation and which stores them in a system configured according to an implementation of the invention. As shown in Figure 4, for a new object, an object creation process generates a new document identifier (docid), which can include a created (created) data and time stamp component and a document identifier component. object (see —above) (402). One or more term producer modules are then summoned to create BA da dasd nx inves names Lama nm tiem dao nb intao (ADA À messaAàsassA by Arnanãsa dh AÁbinta of index 106 (406) and save the object in database 110 (408) In some deployments, each document and term pair can be maintained as a separate entry in a given Index.
    In other deployments, a docid from a docid and term pair can be added to an existing index entry that has the same term.
    For example, an object's document identifier (for example, docid 12345) corresponding to a new Coldplay fan (docid 67890) can be added to one or more existing index entries that have the term "conection.to : 67890 "and / or" conection.fan.to:67890 ". The term producer modules can be updated, and all new new objects will index the new term producer terms.
    In addition, an update process can also regenerate the entire index daily in a MapReduce job so that all old objects are updated with the new terms.
    Index reconstruction can be used as a mechanism to improve performance through denormalization.
    Many storage systems require denormalization of data — at the application level to improve performance.
    The terms producer allows denormalization decisions to be made more dynamically and facilitates changes to. those decisions as changing interrogation patterns.
    Furthermore, changes to: denormalization settings do not require changes to the way in which the underlying data is persistently stored in database 110. For example, assume that a page generation script (home.php) performs the following search frequently to receive status messages from friends: author (conection.from (UID)) type: status.
    If performance becomes a problem due to the volume of interrogation and the size of the type: status posting list, a term producer module can be added or updated so that status messages issue a composite term of type author and mode results to be in a single Index or less posting list.
    For example, a term producer module can be configured to add additional terms for situation objects, such as: ("id": 321224, "type": "status", "created": 1253665137, "message": "author": 12345) A term producer module can be updated to issue the additional term:
    '16/20 so, a set of replica indices can be created for the particular term for further performance enhancements, as described in more detail below. A particular advantage of this scheme is that denormalization decisions can be easily changed, and need not occur at the application level. This means that developers have the ability to store data in the most conceptual logical way. Interrogation performance can be tuned relatively independently of these application-level decisions, rendering applications cleaner, easier to understand, and easier to update over time.
    In a deployment, the search index is fragmented by the document identifier (docid). For example, as Figure 1 illustrates, the index layer can be deployed by hierarchical configuration of index servers including a root server 106a and a plurality of leaf servers 106b. In a deployment, each leaf server 106b is allocated one or more fragments. In another deployment, a group or ring topology can be used. By default, a search can be performed by sending an interrogation for all fragments in parallel, joining the results in the mixer or root index server 106a. In a deployment, a fragment is allocated in one segment. of the document identifier space. In particular modalities, each document identifier (docid) maps (for example, arithmetically or through some mathematical function) to a unique corresponding fragment ID. Consequently, a particular term (for example, "conection.fan.from: 12345") can be kept in a fragment to which the Coldplay object (docid 67890) corresponds and other fragments corresponding to other objects that the individual (docidl2345) established a "fan" connection. In a deployment, each of the 106 index servers is allocated a set of fragment IDs for which they are responsible for maintaining. This allocation can be adjusted to add or remove index servers 106 from the system.
    Sending all queries for all fragments can be computationally expensive and can limit the overall interrogation rate of the system. In a deployment, the index layer deployed by index servers 106 supports special Replica Indexes that index only a subset of the terms in the index system. For example, in addition to a master or master index, the index layer can include one or more additional replica indexes that are adapted for one or more specific types of interrogation. For example, assuming that the conection.from (*) query is an extremely common query in the system. The indexing system described here can be configured so that all terms conection.from are replicated in an additional replica index that contains only those terms. The following command illustrates an interfa- Am gives aAanão Ada anlianãoão AvamPBAlifisadara n1iA narmita 4 arianão da taie (ndirae da replicas = ('"conection.from: *": [...],% Devoted connection replica "email; *; ": [...], & Email search replica" * ": [...],% Main replicas) When an index server 106 performs a query, it chooses the smallest replica index that can satisfy the search. For example, the query connection.from (12345) will be forwarded to an index server that is dedicated to the replica index conection.from. On the other hand, a more generic or broader search, such as —conection.from (12345 ) type: page, will run on the main index or another replica that supports both terms. However, there is no theoretical reason against fragmentation in terms of improving performance for certain queries. One advantage of this model is that the system can support all questions and can be tuned for optimal performance and performance for the most important questions. Once an interrogation becomes common enough, an administrator can tune the system to increase the interrogation rate by creating a dedicated set of replicas. to satisfy that kind of interrogation. This simplifies application development as the network application can, in the first place, be configured to perform any required queries. Before launching the application, replica indices can be created to improve performance based on the structure of the queries created during application development.
    Figure 2 illustrates an exemplary computer system architecture, which can be used to deploy a server 22a, 22b. In one embodiment, the hardware system 1000 comprises a processor 1002, a cache memory 1004, and one or more - modules or executable units, stored in a storage medium and tangible computer, directed to the functions described here. In addition, the 1000 hardware system includes a high performance (1 / O) 1006 input / output bus and a standard 1008 1 / O bus. A 1010 host bridge couples processor 1002 to the 1I bus. / O high performance 1006, while the 1I / O bus bridge 1012 couples two buses 1006 and 1008 to each other. A system memory 1014 and one or more communication / network interfaces 1016 are coupled to bus 1006. Hardware system 1000 may additionally include video memory (not shown) and a display device coupled with video memory. Mass storage 1018, and I / O ports 1020 connect to bus 1008. Hardware system 1000 can optionally include a keyboard and indicating device, and a display device (not mm AStradal ananlada sq barramanta 14NnB0
    '18/20 computer hardware systems, which include, but are not limited to, general-purpose computer systems based on x86-compatible processors manufactured by Intel Corporation of Santa Clara, California, USA, and processors x86 compliant manufactured by Advanced Micro Devices (AMD), Inc., Sunnyvale, California, USA, as well as any other suitable processor.
    The hardware system elements 1000 are described in more detail below. In particular, network interface 1016 provides communication between hardware system 1000 and any one of a wide range of networks, such as an Ethernet network (for example, IEEE 802.3), a background, etc. Mass storage 1018 provides — permanent storage for data and programming instructions to perform the functions described above deployed on servers 22a, 22b, while system memory 1014 (eg DRAM) provides storage temporary for data and programming instructions when executed by processor 1002. I / O ports 620 are one or more serial and / or parallel communication ports that provide communication between additional peripheral devices, which can be coupled to the system of hardware 1000.
    . The hardware system 1000 can include a variety of system architectures; and several hardware system components 1000 can be redeployed. Per . For example, cache 1004 can be on a chip with processor 1002. Alternatively, cache 1004 and processor 1002 can be integrated as a "processor module," with processor 1002 which is called a "processor core." In addition, certain embodiments of the present invention may not require or include all of the above components. For example, the peripheral devices shown coupled to the standard I / O bus 1008 can couple to the high performance I / O bus 1006. In addition, in some embodiments, only a single bus can exist, with the hardware system components 1000 being coupled to the single bus.
    In addition, the hardware system 1000 may include additional components, such as processors, storage devices or additional memories.
    In a deployment, the operations of the modalities described in the present invention are deployed as a series of executable modules operated by the 1000 hardware system, individually or collectively in a distributed computing environment. In a particular modality, a set of modules and / or software units implements a network communications protocol stack, navigation and other computational functions, optimization processes, and the like. The aforementioned functional modules can be made by hardware, executable modules stored in a computer-readable medium
    hardware system, such as processor 1002. Initially, the instruction series can be stored on a storage device, such as mass storage
    1018. However, the series of instructions can be tangibly stored in any suitable storage medium, such as a floppy disk, CD-ROM, ROM, EEPROM, etc. In addition, the instruction series does not need to be stored locally, and could be retrieved from a remote storage device, such as a server on a network, via the 1016 network / communications interface. The instructions are copied from the storage device, such as as mass storage 1018, for memory 1014 and then accessed and executed by processor 1002.
    An operating system manages and controls the operation of hardware system 1000, which includes the input and output of data from software applications (not shown). The operating system provides an interface between the software applications that run on the system and the hardware components of the system. Any suitable operating system can be used, such as the LINUX operating system, the Apple Macintosh operating system, available from Apple Computer Inc. of Cupertino, California, USA, UNIX, Microsoft & WindowsG operating systems, BSD operating systems, and the like .
    . Obviously, other deployments are possible. For example, the nickname generation functions described in the present invention can be implemented in firmware or a circuit. application-specific integrated.
    In addition, the elements and operations described above can be understood from instructions that are stored in storage media. Instructions can be retrieved and executed by a processing system. Some example instructions are software, program code and firmware. Some examples of storage media are memory devices, tape, disks, integrated circuits and servers. The instructions are operational when executed by the processing system to direct the processing system to operate according to the invention. The term "processing system" refers to a single processing device or a group of interoperable processing devices. Some examples of processing devices are integrated circuits and logic circuitry. Those skilled in the art are familiarized with instructions, computers and storage media.
    The present disclosure encompasses all changes, substitutions, variations, alterations, and modifications to the examples of modalities in the present invention that an element of common knowledge in the art would understand. Similarly, where appropriate, the appended claims cover all changes, substitutions, variations, alterations and measures to be taken in accordance with the modalities 12 in connection with the 156 invention in addition to the modalities of the present invention have been described as operational in conjunction with a website. social networking web, the present invention can be used in conjunction with any communications mode that supports web applications and model data as an association chart.
    Additionally, in some embodiments, the term "web service" and "website daweb" can be used interchangeably and additionally can refer to a common or generalized API on a device, such as a mobile device (for example, cell phone , smart phone, personal GPS, personal digital assistant,
    positive personal game, etc.), which makes API calls directly to a server.
    权利要求:
    Claims (8)
    [1]
    1. Apparatus FEATURED for understanding: an Index stored in one or more memory devices, the index comprising: a plurality of index entries, where each entry comprises a term and a document identifier, at least one of the entry entries index comprising: a document identifier corresponding to a data object from a plurality of data objects; and a composite term comprising two or more data object attributes; and one or more operational index servers to: access a query that comprises one or more attribute values; search the Index against attribute values to identify one or more data objects that correspond to the query; and return one or more object identifiers corresponding to one or more identified data objects that correspond to the query and an operational term producer to: access a data object comprising a plurality of data object attributes; analyze the data object to create one or more index entries for the data object, and produce a set of pairs of document identifier terms, each pair comprising a type based on object types corresponding to the data objects, where at least one of the terms is a composite term, pairs of document identifying terms comprising: a type of connection and a source pair comprising: a document identifier of a source object; and a composite term that identifies a connection type for an edge object and a document identifier for a target object, a source and connection pair comprising: a document identifier for the source object; and a composite term that identifies a connection and a document identifier for a target object; a type of connection and destination pair comprising: a document identifier of the destination object; and a composite term that identifies a connection type for an edge object and a document identifier for a source object, or a destination and connection pair that comprises:
    a document identifier for a target object; and a composite term that identifies a connection and document identifier for a source object.
    [2]
    2. Apparatus, according to claim 1, CHARACTERIZED by the fact that the document identifier comprises an object identifier and a time stamp.
    [3]
    3. Apparatus, according to claim 1, CHARACTERIZED by the fact that the Index comprises one or more replica indices.
    [4]
    4. Apparatus, according to claim 1, CHARACTERIZED by the fact that the index comprises a plurality of fragments, each being allocated in a portion of a document identifier space.
    [5]
    5. Method CHARACTERIZED for understanding: storing, in one or more memory devices, an index comprising: a plurality of index entries, where each entry comprises a term and a document identifier; at least one of the index entries comprising: a document identifier corresponding to a data object from a plurality of data objects; and a composite term comprising two or more attributes of data objects; and accessing, through one or more index servers, a query that comprises one or more attribute values; search, through one or more index servers, the index against the attribute values, to identify one or more identified data objects that correspond to the query; return, through the index servers, one or more object identifiers corresponding to one or more identified data objects that correspond to the query. access, through one or more term producers, a data object comprising a plurality of data object attributes; analyze, through term producers, a data object to create one or more index entries for the data object; and produce, through term producers, a set of pairs of document-identifying terms, each pair comprising a type based on corresponding object types of data objects, where at least one of the terms is a term with - purpose, document-term identifier pairs comprising: a type of connection and a source pair comprising: a document identifier of a source object and a composite term that identifies a type of connection of an edge object and a document identifier for a target object, a source and connection pair comprising: a document identifier for a source object; and a composite term that identifies a connection and a document identifier for a target object; a type of connection and destination pair comprising: a document identifier for a destination object; and a composite term that identifies a connection type for an edge object and a document identifier for a source object, or a destination and connection pair that comprises: a document identifier for a destination object; and a composite term that identifies a connection and document identifier for a source object.
    [6]
    6. Method, according to claim 5, CHARACTERIZED by the fact that the document identifier comprises an object identifier and a time stamp.
    [7]
    7. Method, according to claim 5, CHARACTERIZED by the fact that the Index comprises one or more replica indices.
    [8]
    8. Method, according to claim 5, CHARACTERIZED by the fact that the Index comprises a plurality of fragments, each allocated in a portion of a document identifier space.
    类似技术:
    公开号 | 公开日 | 专利标题
    BR112013016926A2|2020-10-27|composite term index for graphic data
    US9886484B2|2018-02-06|Distributed cache for graph data
    AU2013270511B2|2015-11-26|Composite term index for graph data
    同族专利:
    公开号 | 公开日
    KR101419828B1|2014-07-16|
    EP2659402A4|2016-08-03|
    MX2013007685A|2013-12-02|
    AU2011353034B2|2013-10-10|
    JP2017211999A|2017-11-30|
    US8527497B2|2013-09-03|
    JP2014096164A|2014-05-22|
    EP2659402A1|2013-11-06|
    AU2011353034A1|2013-07-18|
    US9576060B2|2017-02-21|
    AU2017201389A1|2017-03-23|
    CN107092666A|2017-08-25|
    CN103348344B|2017-05-03|
    MX350594B|2017-09-11|
    AU2016200896B2|2016-11-24|
    US20120215785A1|2012-08-23|
    CN103348344A|2013-10-09|
    US20140006412A1|2014-01-02|
    CA2848100C|2016-07-12|
    KR20130096767A|2013-08-30|
    JP2014507706A|2014-03-27|
    US9934329B2|2018-04-03|
    JP5451954B1|2014-03-26|
    CA2928937A1|2012-07-05|
    JP5964808B2|2016-08-03|
    US9223899B2|2015-12-29|
    JP6174762B2|2017-08-02|
    WO2012091844A1|2012-07-05|
    AU2016200896A1|2016-03-03|
    CA2928937C|2017-08-15|
    EP2659402B1|2021-01-06|
    CN107092666B|2018-06-15|
    JP2016189214A|2016-11-04|
    CA2848100A1|2012-07-05|
    JP6326532B2|2018-05-16|
    US20170154124A1|2017-06-01|
    CA2823146A1|2012-07-05|
    CA2972316C|2018-10-16|
    MX355952B|2018-05-07|
    US20160048600A1|2016-02-18|
    CA2823146C|2014-07-15|
    CA2972316A1|2012-07-05|
    AU2017201389B2|2018-03-08|
    引用文献:
    公开号 | 申请日 | 公开日 | 申请人 | 专利标题
    JPH05313971A|1992-05-13|1993-11-26|Mitsubishi Electric Corp|Key word managing system in relational data base|
    JP3132606B2|1993-02-25|2001-02-05|日本電気株式会社|Related information converter|
    US5524240A|1994-05-24|1996-06-04|Panasonic Technologies, Inc.|Method and apparatus for storage and retrieval of handwritten information|
    JPH0944521A|1995-07-25|1997-02-14|Fuji Xerox Co Ltd|Index generating device and document retrieval device|
    JPH09198395A|1996-01-16|1997-07-31|Fuji Xerox Co Ltd|Document retrieval device|
    JP2965018B2|1997-11-17|1999-10-18|日本電気株式会社|Search information display method and search information display device in hypermedia system|
    JP3347088B2|1999-02-12|2002-11-20|インターナショナル・ビジネス・マシーンズ・コーポレーション|Related information search method and system|
    JP4042295B2|2000-04-14|2008-02-06|富士ゼロックス株式会社|Keyword extraction rule refinement support device, storage medium, and keyword extraction rule refinement support method|
    WO2002084431A2|2001-04-11|2002-10-24|International Business Machines Corporation|Simplifying and manipulating k-partite graphs|
    US20040024755A1|2002-08-05|2004-02-05|Rickard John Terrell|System and method for indexing non-textual data|
    JP2004272831A|2003-03-12|2004-09-30|Ricoh Co Ltd|Keyword extracting device, program and recording medium|
    KR20060011828A|2003-04-03|2006-02-03|더블린 시티 유니버시티|Shape matching method for indexing and retrieving multimedia data|
    CN1826598A|2003-07-21|2006-08-30|皇家飞利浦电子股份有限公司|Method of searching in a collection of documents|
    US7779345B2|2003-07-30|2010-08-17|Aol Inc.|Reverse mapping method and apparatus for form filling|
    JP2005210384A|2004-01-22|2005-08-04|Konica Minolta Medical & Graphic Inc|Image processing method, image processor, and image processing program|
    JP2006185412A|2004-12-03|2006-07-13|Kazusa Dna Kenkyusho|Information processor, information processing method and program thereof|
    US7526486B2|2006-05-22|2009-04-28|Initiate Systems, Inc.|Method and system for indexing information about entities with respect to hierarchies|
    US7752243B2|2006-06-06|2010-07-06|University Of Regina|Method and apparatus for construction and use of concept knowledge base|
    CN101558331B|2006-12-14|2011-07-27|松下电器产业株式会社|Lens and method for manufacturing the same|
    US7702620B2|2007-03-29|2010-04-20|International Business Machines Corporation|System and method for ranked keyword search on graphs|
    US8166021B1|2007-03-30|2012-04-24|Google Inc.|Query phrasification|
    US7693813B1|2007-03-30|2010-04-06|Google Inc.|Index server architecture using tiered and sharded phrase posting lists|
    US8166045B1|2007-03-30|2012-04-24|Google Inc.|Phrase extraction using subphrase scoring|
    US8086594B1|2007-03-30|2011-12-27|Google Inc.|Bifurcated document relevance scoring|
    JP4930153B2|2007-03-30|2012-05-16|富士通株式会社|Document search system, document number subsequence acquisition apparatus, and document search method|
    US7925655B1|2007-03-30|2011-04-12|Google Inc.|Query scheduling using hierarchical tiers of index servers|
    US7702614B1|2007-03-30|2010-04-20|Google Inc.|Index updating using segment swapping|
    US7904461B2|2007-05-01|2011-03-08|Google Inc.|Advertiser and user association|
    JP5075653B2|2008-01-21|2012-11-21|株式会社日立製作所|Database management method, database management apparatus, database management program, and database system|
    JP2009176044A|2008-01-24|2009-08-06|Nippon Telegr & Teleph Corp <Ntt>|Document retrieving method, device, program, and computer readable recording medium|
    US10185779B2|2008-03-03|2019-01-22|Oath Inc.|Mechanisms for content aggregation, syndication, sharing, and updating|
    KR100936198B1|2008-03-21|2010-01-11|인하대학교 산학협력단|Social network analysing system|
    CN101770459B|2008-12-31|2013-10-09|华为技术有限公司|Method and device for commending friends|
    US8438173B2|2009-01-09|2013-05-07|Microsoft Corporation|Indexing and querying data stores using concatenated terms|
    WO2010082207A1|2009-01-16|2010-07-22|Sanjiv Agarwal|Dynamic indexing while authoring|
    US20100198844A1|2009-01-30|2010-08-05|Thales-Raytheon Systems Company, LLC|Automatic Code Generation|
    WO2010101540A1|2009-03-02|2010-09-10|Panchenko Borys Evgenijovich|Method for the fully modifiable framework distribution of data in a data warehouse taking account of the preliminary etymological separation of said data|
    US20100241580A1|2009-03-19|2010-09-23|Tagged, Inc.|System and method of selecting a relevant user for introduction to a user in an online environment|
    US8185544B2|2009-04-08|2012-05-22|Google Inc.|Generating improved document classification data using historical search results|
    US8316056B2|2009-12-08|2012-11-20|Facebook, Inc.|Second-order connection search in a social networking system|
    US8180804B1|2010-04-19|2012-05-15|Facebook, Inc.|Dynamically generating recommendations based on social graph information|
    US8572129B1|2010-04-19|2013-10-29|Facebook, Inc.|Automatically generating nodes and edges in an integrated social graph|
    US8185558B1|2010-04-19|2012-05-22|Facebook, Inc.|Automatically generating nodes and edges in an integrated social graph|
    US8527497B2|2010-12-30|2013-09-03|Facebook, Inc.|Composite term index for graph data|
    US8832111B2|2010-12-30|2014-09-09|Facebook, Inc.|Distributed cache for graph data|US8832111B2|2010-12-30|2014-09-09|Facebook, Inc.|Distributed cache for graph data|
    US8527497B2|2010-12-30|2013-09-03|Facebook, Inc.|Composite term index for graph data|
    US8863153B2|2011-09-13|2014-10-14|Sap Se|Situational recommendations in heterogenous system environment|
    KR20130049111A|2011-11-03|2013-05-13|한국전자통신연구원|Forensic index method and apparatus by distributed processing|
    US8700673B2|2012-02-16|2014-04-15|Oracle International Corporation|Mechanisms for metadata search in enterprise applications|
    US8682932B2|2012-02-16|2014-03-25|Oracle International Corporation|Mechanisms for searching enterprise data graphs|
    US8484208B1|2012-02-16|2013-07-09|Oracle International Corporation|Displaying results of keyword search over enterprise data|
    US8935255B2|2012-07-27|2015-01-13|Facebook, Inc.|Social static ranking for search|
    US8825711B2|2012-08-13|2014-09-02|Magnet Systems, Inc.|Managing cross-correlated data|
    JP5959405B2|2012-10-30|2016-08-02|株式会社Uacj|Aluminum alloy foil|
    US9576007B1|2012-12-21|2017-02-21|Google Inc.|Index and query serving for low latency search of large graphs|
    US8935271B2|2012-12-21|2015-01-13|Facebook, Inc.|Extract operator|
    US9659043B2|2013-02-07|2017-05-23|Enigma Technologies, Inc.|Data system and method|
    US9171063B2|2013-03-13|2015-10-27|Facebook, Inc.|Short-term hashes|
    US9712608B2|2013-03-14|2017-07-18|Microsoft Technology Licensing, Llc|Elastically scalable document-oriented storage services|
    US9910887B2|2013-04-25|2018-03-06|Facebook, Inc.|Variable search query vertical access|
    US9223898B2|2013-05-08|2015-12-29|Facebook, Inc.|Filtering suggested structured queries on online social networks|
    US10983973B2|2013-05-21|2021-04-20|Facebook, Inc.|Database sharding with incorporated updates|
    US10977229B2|2013-05-21|2021-04-13|Facebook, Inc.|Database sharding with update layer|
    US9483568B1|2013-06-05|2016-11-01|Google Inc.|Indexing system|
    CN104714952B|2013-12-11|2018-07-06|航天信息股份有限公司|A kind of multi-retrieving method of database|
    CN104714973A|2013-12-17|2015-06-17|航天信息股份有限公司|Method and device for processing query statements|
    US9652554B2|2013-12-26|2017-05-16|Facebook, Inc.|Systems and methods for adding users to a networked computer system|
    US8954441B1|2014-01-02|2015-02-10|Linkedin Corporation|Graph-based system and method of information storage and retrieval|
    US20150229715A1|2014-02-13|2015-08-13|Linkedin Corporation|Cluster management|
    US9665633B2|2014-02-19|2017-05-30|Snowflake Computing, Inc.|Data management systems and methods|
    CN103810300B|2014-03-10|2017-08-01|北京国双科技有限公司|The data query method and apparatus covered for non-index|
    US9514123B2|2014-08-21|2016-12-06|Dropbox, Inc.|Multi-user search system with methodology for instant indexing|
    US9384226B1|2015-01-30|2016-07-05|Dropbox, Inc.|Personal content item searching system and method|
    US9183303B1|2015-01-30|2015-11-10|Dropbox, Inc.|Personal content item searching system and method|
    US10037388B2|2015-04-27|2018-07-31|Microsoft Technology Licensing, Llc|Fast querying of social network data|
    US10127230B2|2015-05-01|2018-11-13|Microsoft Technology Licensing, Llc|Dynamic content suggestion in sparse traffic environment|
    US10007879B2|2015-05-27|2018-06-26|International Business Machines Corporation|Authoring system for assembling clinical knowledge|
    US9514247B1|2015-10-28|2016-12-06|Linkedin Corporation|Message passing in a distributed graph database|
    US20170228409A1|2016-02-08|2017-08-10|Red Hat, Inc.|In-memory journaling|
    US10180992B2|2016-03-01|2019-01-15|Microsoft Technology Licensing, Llc|Atomic updating of graph database index structures|
    US10599656B1|2016-03-04|2020-03-24|Twitter, Inc.|Indexing and data storage for realtime and contemporaneous content suggestions|
    CN105930506A|2016-05-10|2016-09-07|乐视控股（北京）有限公司|Object retrieval method and object storage system|
    US10540332B2|2016-08-03|2020-01-21|Microsoft Technology Licensing, Llc|Efficient denormalization of data instances|
    US10366368B2|2016-09-22|2019-07-30|Microsoft Technology Licensing, Llc|Search prioritization among users in communication platforms|
    US10754859B2|2016-10-28|2020-08-25|Microsoft Technology Licensing, Llc|Encoding edges in graph databases|
    CN108093020B|2016-11-23|2020-12-18|阿里巴巴集团控股有限公司|Data processing method and device|
    US10445321B2|2017-02-21|2019-10-15|Microsoft Technology Licensing, Llc|Multi-tenant distribution of graph database caches|
    US10565197B2|2017-03-02|2020-02-18|International Business Machines Corporation|Search performance using smart bitmap operations|
    US10671671B2|2017-06-09|2020-06-02|Microsoft Technology Licensing, Llc|Supporting tuples in log-based representations of graph databases|
    US10445370B2|2017-06-09|2019-10-15|Microsoft Technology Licensing, Llc|Compound indexes for graph databases|
    US10628492B2|2017-07-20|2020-04-21|Microsoft Technology Licensing, Llc|Distributed graph database writes|
    WO2019055549A1|2017-09-12|2019-03-21|Schlumberger Technology Corporation|Data searching, enrichment and consumption techniques using exploration and/or production entity relationships|
    US10963514B2|2017-11-30|2021-03-30|Facebook, Inc.|Using related mentions to enhance link probability on online social networks|
    US10129705B1|2017-12-11|2018-11-13|Facebook, Inc.|Location prediction using wireless signals on online social networks|
    US10983997B2|2018-03-28|2021-04-20|Microsoft Technology Licensing, Llc|Path query evaluation in graph databases|
    CN108549644A|2018-04-12|2018-09-18|苏州大学|Omission pronominal translation method towards neural machine translation|
    US11030242B1|2018-10-15|2021-06-08|Rockset, Inc.|Indexing and querying semi-structured documents using a key-value store|
    CN109063215B|2018-10-16|2020-10-30|成都四方伟业软件股份有限公司|Data retrieval method and device|
    US11113267B2|2019-09-30|2021-09-07|Microsoft Technology Licensing, Llc|Enforcing path consistency in graph database path query evaluation|
    法律状态:
    2020-11-10| B06U| Preliminary requirement: requests with searches performed by other patent offices: procedure suspended [chapter 6.21 patent gazette]|
    2021-02-23| B11B| Dismissal acc. art. 36, par 1 of ipl - no reply within 90 days to fullfil the necessary requirements|
    2021-12-07| B350| Update of information on the portal [chapter 15.35 patent gazette]|
    优先权:
    申请号 | 申请日 | 专利标题
    US201061428615P| true| 2010-12-30|2010-12-30|
    US61/428.615|2010-12-30|
    US13/228,312|US8527497B2|2010-12-30|2011-09-08|Composite term index for graph data|
    US13/228.312|2011-09-08|
    PCT/US2011/062603|WO2012091844A1|2010-12-30|2011-11-30|Composite term index for graph data|
    [返回顶部]