CN109446358A - A kind of chart database accelerator and method based on ID caching technology - Google Patents
A kind of chart database accelerator and method based on ID caching technology Download PDFInfo
- Publication number
- CN109446358A CN109446358A CN201810981901.1A CN201810981901A CN109446358A CN 109446358 A CN109446358 A CN 109446358A CN 201810981901 A CN201810981901 A CN 201810981901A CN 109446358 A CN109446358 A CN 109446358A
- Authority
- CN
- China
- Prior art keywords
- caching
- sentence
- data
- database
- query
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses the cores of chart database ID caching technology, applied to the caching on the chart database of million grades of above data amounts with dynamic, ID caching technology has cached ID in memory, although still to inquire chart database, but it is huge to have raised speed compared to the query process for originally traversing data up to a million, it only needs to be oriented inquiry according to ID, queries is very small, therefore inquiry velocity is also promoted huge, and there is data dynamic, data variation in database does not conflict with caching, significantly reduces dirty data probability.It is divided into following components: the special buffer structure formulated based on chart database structure feature, Different matching strategy towards simple queries and complex query, client is notified to be updated caching when database has change to reach consistency, the Replacement Strategy that caching quantity carries out when excessive.
Description
Technical field
The invention belongs to computer graphic database technical field more particularly to the debugging technique fields of chart database software.
Background technique
Euler is played in chart database source and Tu is theoretical, and alternatively referred to as towards the/database based on figure, corresponding English is
Graph Database.The basic meaning of chart database is with " figure " this data structure storage and to inquire data, rather than deposit
Store up the database of picture.Its data model can also handle key-value pair mainly with node and relationship (side) Lai Tixian.It excellent
Point is the quick relations problems for solving complexity.
Computer storage wishes to store relational data, although relevant database can accomplish this point, but
It does badly, results in data redundancy to store relationship, and be unable to the dynamic of conformity relation data, it can not be good
Support the multilayer complex query of relation data.Internal relation between data aiming at the problem that complicated and dynamic change, come into being
It is chart database, it is known that chart database is one of future directions of database purchase, by memory node and relationship come real
The now abstract storage of figure, it has also become the preference data library of social field, financial field and traffic information field etc., but with number
According to the growth of scale and data complexity, the inquiry velocity of chart database is unable to satisfy current demand, it is even more impossible to relationship type
Database is compared, and the introducing at this point for caching technology is very important.
Current chart database has neo4j, titan, arangodb, orientdb, gun etc..On caching technology
Orientdb is file cache cache, and file buffering cache has cached the file data of same format.But it is frequent due to inquiring
It is mutually inquired between node and relationship, and relationship and node are placed on the file data of different-format, so caching speed is not
Fastly.Neo4j is subquery cache, hereof by the subquery cache inquired, but due to caching be data result, so caching
Shared data volume is very big.And it is safe it is smooth used multi-layer data caching, with graph traversal easily and fast.They are smooth from a Thailand
It is accessed in affairs.It caches closer to affairs, the speed of cache access is faster, and EMS memory occupation and maintenance costs are also higher.
It is very long that inquiry simple data time-consuming is frequently encountered when we are using chart database, or even is being looked into sometimes
Asking complex data will appear the case where can not returning the result.The main reason is that the storage mode and storage organization of chart database are led
Cause data query very slow, for example certain chart database possesses 1,000,000 nodes, 2,000,000 relationships.One simple use
Family relationship slow nearly several times of the inquiry velocity compared with relevant database in same hardware environment, this can not put up with,
Main cause is because inquiry requires to toggle matching between relationship and node each time.But this is not that can not solve
The problem of, the place different from relevant database is that chart database is stored with node and relationship, node and relationship
The ID for respectively corresponding a global assignment, unlike the ID distribution system of relevant database each table, that is, different.If it is known that
The ID number of the query result, can direct locating query, traverse full document query without going again.
It there has been the caching technology based on ID inquiry mechanism in summary, then why not direct as tradition caching
The data object of query result is stored, mainly there are several reasons: storing data is the result is that very huge, some inquiry knots
Fruit may contain hundreds and thousands of a nodes, this if only saves ID number (Long class to the just huge consumption of memory
Type) expense is very small.Secondly, generating cache table dirty data, storing data result will be unable to carry out data with chart database
It unanimously, is also very big to reach consistent computer resource expense, so then at least ensure that diagram data using ID
Library content is consistent, and serious forgiveness is low.
Summary of the invention
The purpose of the present invention is to solve the I O access number mistakes that the prior art occurs in chart database query process
Mostly with slow-paced serious problems, a kind of chart database accelerator and method based on ID caching technology is proposed.
To achieve the goals above, technical solution of the present invention:
A kind of chart database device based on ID caching technology, including
Server is accessed, the query statement and ID caching accelerator for obtaining user's input communicate, and feedback is looked into
Result is ask to user.
Chart database, for storing the interface of the data for needing to inquire and open ID inquiry.
ID caches accelerator, for parsing query statement, the corresponding result ID of inquiry query statement, with chart database into
Row communication, and feedback query result gives access server.
ID caches mapping table: the corresponding result ID of sentence for storing historical query, and possesses three layers of tree rope
Draw.
Caching query module: the query statement for inputting to user carries out being parsed into simple statement or complicated sentence,
Three layer analysis are carried out if it is simple statement, is then indexed and matches its corresponding result ID.If it is complicated sentence, directly
The full word section of query statement is matched, it is then obtained and corresponds to result ID.
Buffer consistency module: for updating caching, the result ID and chart database data consistency of caching are realized.
Caching replacement module: for eliminating useless caching, when deleting data cached low frequency and data cached spilling
Carry out caching reduction.
Wherein, the access server uses http to communicate with the communication of accelerator.
Wherein, the thing TCP full-duplex communication that communication of the accelerator with chart database uses, it is therefore an objective to when carrying out long
Between access, reduce connect consumed by the time.
Wherein, the open ID query interface of the chart database is inquiry mechanism always specific to chart database itself.
Wherein, three layers of tree is being embodied for chart database storage organization, main contents are as follows: first is class
Type is divided into node type and relationship type, and the second layer is Label data, is the customized name of user, is equivalent to relationship type number
Show that third layer is condition according to library, is the core content of user query sentence, querying condition.
Wherein, whether the simple statement and complicated sentence refer to comprising with, the inquiry of these multilists of union etc.
Sentence is the simple statement of single table inquiry if not, and if so, being the complicated sentence that multilist is even looked into.
Wherein, the very high inquiry of once frequency that the useless caching refers to, but subsequent a very long time does not use
It crosses.
Wherein, the low frequency caching refers to sporadic inquiry, only accesses several times in one week, this subquery cache
Value is too low, and influences matching speed.
Wherein, the cache overflow refers to occur newly to delay when caching maximum storage quantity obtains when accelerator has
Loss, that is, so-called spilling are deposited, while will affect the other function of device.
A kind of implementation method that the chart database based on ID caching technology accelerates, includes the following steps:
ID caches the query statement of mapping block storage history, and establishes three-layer indexing structure and mapping result ID.
User inputs query sentence:
Query statement, which enters access server and communicated by http, to be passed to ID buffer storage and is matched.
By input by sentence to caching query module after arrival ID buffer storage, corresponding knot is inquired from ID caching mapping table
Then result ID is re-assemblied query statement as querying condition and passes through TCP communication calling figure database interface by fruit ID.
The speciality of chart database is that it possesses 2 globally consistent ID distribution mechanisms, that is to say, that node and relationship without
By being any label, ID does not have conflict, at the same also imply that this ID number be it is globally unique, can be used as look into completely
Inquiry condition.Therefore inquiry ID will not be related to multiple I O read file, it is thus only necessary to read file where ID.
Then chart database can return the result after receiving query statement according to ID with quick indexing to corresponding query result.
User input database updata sentence:
Change sentence enters access server and passes to ID buffer storage progress consistency by http communication.
By input by sentence to buffer consistency module after arrival ID buffer storage, more newspeak is inquired from ID caching mapping table
Sentence institute is relevant data cached, then carries out buffer update.
Then sentence chart database is transmitted to execute.
Data cached arrival spill point:
Caching replacement module is executed, the cache priority that low and medium frequency uses will be cached and replaced away.If overflowing situation does not have
Improve and execute two wheel displacements, will finally be gone out using time earliest caching replacement in caching.
Wherein, the three-layer indexing structure and mapping result ID are one-to-one, and three layer indexs are type, label respectively
And querying condition, it can be from a query statement fractionation, as a result ID is then pair of the historical query result of this sentence
As ID.
Wherein, the matching process refers to that it is type, label and inquiry item that a query statement, which is carried out key word split,
Part.Then it is corresponded with ID mapping table, and if so, returning the result ID carries out database ID inquiry, if there is no
Then return cache adds.
Wherein, the consistency process refers to that it is type, label that a update sentence is carried out to key word split.Then
It is corresponded with ID cache table, and if so, thinking that the caching will be updated is an expired caching, is directly deleted
Then it is executed in chart database and updates sentence, if there is no then directly execution.
Wherein, the spill point is default value, is set according to system performance and memory size.
Wherein, when the low frequency refers to that the caching is queried to and returns the result ID each time, can all there is one
A count attribute carries out adding one, so thinking that frequency of use is extremely low if the numerical value is very low, can preferentially delete.
Beneficial effects of the present invention: apparatus and method of the present invention among server and chart database by adding one
ID buffer storage, it is corresponding for by the analysis to inquiring and updating sentence, it to be carried out with database ID, use consistency mould
Block guarantees that caching is consistent with chart database, and is accelerated using the distinctive ID inquiry mechanism of chart database, original not changing
Any configuration of chart database and code situation, which realize, only uses ID inquiry mechanism, and improves inquiry velocity and reduce IO reading
Take number.
Detailed description of the invention
Fig. 1 is a kind of system structure diagram of chart database accelerator based on ID caching technology of the invention.
Fig. 2 is a kind of modular structure schematic diagram of chart database accelerated method based on ID caching technology of the invention.
Fig. 3 is the detail flowchart of the matching module of the method for the present invention.
Fig. 4 is the detail flowchart of the consistency module of the method for the present invention.
Fig. 5 is the detail flowchart of the replacement module of the method for the present invention.
Specific embodiment
Understand in order to make the purpose of the present invention, the technical scheme and advantages are more clear, the present invention is done below in conjunction with attached drawing
It further illustrates.
As shown in Figure 1, a kind of system structure diagram of the chart database accelerator based on ID caching technology, You Tuke
Know including at least one server, at least one chart database and at least one ID cache accelerator, and the server is external
Open interface, the chart database are opened to the outside world ID query interface, which is adapted to major application layer, including
Android, IOS and JS etc..It is also adapted to major operating system simultaneously, including windows, liunx and aliyun etc..
As shown in Fig. 2, a kind of modular structure schematic diagram of the chart database accelerated method based on ID caching technology, including such as
Lower module:
The query sentence of database of input is matched to corresponding caching ID by cache match module.
Buffer consistency module, by the sentence to database manipulation analyze then simultaneously more new database ease up deposit,
To reach consistency.
ID caches mapping table, a mapping table being stored in memory in the device, and key-value pair is result ID respectively and looks into
Ask the analysis structure of sentence.
Caching replacement module, data volume reach as defined in spill point when, caching is replaced, exclude it is useless and
Inefficient caching, has reached Buffer Utilization maximum.
As shown in figure 3, the detail flowchart of the matching module of the method for the present invention, query statement and caching number are illustrated in figure
According to matching process, query statement is parsed into three-decker sentence in the matching process, is then corresponded with caching,
And result ID is finally returned to, detailed process is explained as follows:
When a query statement reaches device, the first step can enter simple complicated sentence judgement, if it is simple
Sentence then carries out three-decker analysis, and query statement is parsed into 3 layers of fixed format, such as Match (u:User) where
U.name=' mary ' return u.It is node that sentence energy study plot, which analyzes type, and label object is User and querying condition
For name=' mary '.Also have three-decker in mapping table simultaneously, is corresponded with mapping table data, and if so,
Return the result, if there is no be then recorded caching in, it is data cached as what is inquired again next time, with accelerate inquiry knot
Fruit.
If it is complex data then do not have to carry out three-decker analysis because complex data is excessively complicated, three layer analysis without
Method carries out, so carrying out full word piecewise analysis, such as Match (u:User) where u.name=' mary ' with u Match
(u1:User) where u.age=u1.age with u1Match (u2:User) where u1.name=u2.name
The sentence that return u2. can not be analyzed here it is one, so we will carry out in its full word segment and complicated cache table
Matching, returns the result if be matched to, and is recorded in caching if mismatching, as the caching number inquired again next time
According to.
As shown in figure 4, the detail flowchart of the consistency module of present aspect method, carries out the main basis of consistency algorithm
Invalidation protocols technology notifies client to be updated caching to reach strongly consistent when database has change
Property, for having the data of change, such as update and delete operation in database, its ID is recorded, buffer update algorithm is carried out.For
There are newly-increased data in database, similitude comparison will be carried out, similitude comparison is for newly-increased data, in the mark of same type
Each attribute is inquired in label, there is the data record ID containing same alike result, it is believed that is high probability set of metadata of similar data, is carried out buffer update calculation
Method.It illustrates to update sentence and data cached consistency process in figure, query statement is parsed into three during consistency
Layer structured statement, is then corresponded with caching, and if so, cancel statement is to reach database update, detailed process
It is explained as follows:
When a update sentence reaches device, three layer analysis are carried out to sentence is updated, then to caching mapping table
In data matched, if successful match, take out relevant ID, then according to correlation ID again to caching mapping table in
Result ID matched, deleted as long as successful match, such as Match (u:User) where u.name=' tom ' set
U.age=14.The sentence indicates that the age by the people that name is tom is all changed to 14, and tom this data ID is assumed to be
12, then we will include 12 data all deletions result ID all inside cache table according to matching result.
As shown in figure 5, the detail flowchart of the replacement module of the method for the present invention, each has pair when being buffered in addition
The creation time answered uses time and number recently, and according to Time-to-live fields technology, and not all caching is all
High frequency, for example log in once, browser records its Entered state possible one month or even will not all carry out again some months identical
Inquiry.For sporadic query statement, should swap out in time optimal to reach caching.Therefore following algorithm is formulated in 24 hours
Sporadic sentence is considered without the sentence of 3 access times, does not cache necessity, for being lower than 7 access times in one week
Sentence be considered low recurrence sentence, deposit preferentially swaps out array.When caching quantity is excessive or memory will overflow, execute
Inefficient caching removes algorithm, the sentence preferentially to swap out in array is swapped out, while will use the time recently according to according to LFU strategy
Most remote swaps out.Detailed process is explained as follows:
When data cached quantity spill point on earth, we will delete the low-down caching number of frequency of use first
According to continuing to delete last time using time earliest data, and be finally reached if alleviation has not been reached yet
Caching is alleviated.
Those of ordinary skill in the art will understand that the embodiments described herein, which is to help reader, understands this hair
Bright principle, it should be understood that protection scope of the present invention is not limited to such specific embodiments and embodiments.This field
Those of ordinary skill disclosed the technical disclosures can make according to the present invention and various not depart from the other each of essence of the invention
The specific variations and combinations of kind, these variations and combinations are still within the scope of the present invention.
Claims (6)
1. chart database accelerator and method based on ID caching technology, which is characterized in that be made of five modules, respectively
It is:
ID caches accelerator, and for parsing query statement, the corresponding result ID of inquiry query statement is led to chart database
Letter, and feedback query result gives access server.
ID caches mapping table: the corresponding result ID of sentence for storing historical query, and possesses three layers of tree index.
Caching query module: the query statement for inputting to user carries out being parsed into simple statement or complicated sentence, if
Three layer analysis are carried out for simple statement, is then indexed and matches its corresponding result ID.If it is complicated sentence, directly to looking into
The full word section for asking sentence is matched, and is then obtained it and is corresponded to result ID.
Buffer consistency module: for updating caching, the result ID and chart database data consistency of caching are realized.
Caching replacement module: it for eliminating useless caching, is carried out when deleting data cached low frequency and data cached spilling
Caching reduction.
Detailed process is as follows for it:
ID query caching technology is between chart database and query interface, when having to the inquiry operation of chart database from answering
Before being transmitted to interface calling database with layer, we will carry out ID cache match, go to inquiry operation if successful match
ID is inquired above, by data in the ID quick search to database of caching and is returned, and is called if being not matched to normal
Data base querying, while the record buffer memory after returning the result.When there is the operation for carrying out data modification to database, then in data
ID is returned after the modification of library and carries out consistency algorithm, deletes matched ID in caching.It is equipped with simultaneously and caches expired replacement algorithm.
2. parsing sentence as described in claim 1, which is characterized in that the strategy for parsing sentence is as follows with basic ideas:
It is required that one, sentence is whether to carry complicated statement keyword, such as with, and the complex data base operations sentence such as unit contains
Have then for complicated sentence, not containing then is simple statement；
It is required that two, simple statement carries out three-decker analysis, and three-decker is type, label and querying condition respectively.
3. ID as described in claim 1 caches mapping table, which is characterized in that it is as follows that ID caches mapping table main contents:
It includes simple cache mapping and complicated caching mapping, simple cache mapping storage three-decker data carry out pointer rope
Draw, accelerate inquiry, result is database ID, this is because possessing Global ID the characteristics of chart database, storage ID can reduce slow
The capacity deposited increases performance, while guaranteeing the consistency of database and caching.Three-decker is type, label and inquiry respectively
Condition, it is current so simple statement is all made of these three, as long as so simple statement can analyze this three layers of title,
And acceleration can be indexed in inquiry by being parsed into three-decker.Another kind is complicated caching mapping table, since complicated sentence is multiple
It is miscellaneous it is changeable can not carry out being parsed into three layers, so the sentence of full word section can only be stored directly, inquiry when, can only also inquire complete
Field.
4. caching query module as described in claim 1, which is characterized in that caching looks into the main policies reflected and content is as follows:
Data base manipulation statement input, analyzes it and carries out three if it is simple statement for complicated or simple statement
The number of plies it is judged that, such as Match (u:User) where u.name=' mary ' return u.The sentence energy study plot analyzes
Type is node, and label object is User and querying condition is name=' mary '.With the simple cache table progress in memory
Match, no longer progress data base querying is returned the result if containing, then updates caching if carrying out data base querying without if
Table.Simple cache matching is to first determine whether that inquiry whether containing non-simple queries languages such as with, uses parsing side if not containing
Method carries out query statement to parse type, three fields of label and querying condition, then first by the class of type and ID mapping table
Type-word section is matched, and is again matched label with label under the type object after successful match, will be looked into again after successful match
The query statement array object ask in the sentence label object is matched.It is arranged whether judgement finally hits simultaneously, no hit
Then carry out conventional query and addition caching.
The inquiry of full word section, such as Match (u:User) where u.name=' mary ' are then directly carried out if it is complicated sentence
With u Match (u1:User) where u.age=u1.age with u1 Match (u2:User) where u1.name
The sentence that=u2.name return u2. can not be analyzed here it is one, so we delay its full word segment with complexity
It deposits in table and is matched, ID is returned if successful match, data base querying is carried out if it fails to match and update memory ID and is slow
It deposits.Complicated cache match is to first determine whether query statement is containing the non-simple queries language such as with, if containing thinking for not
Sentence can be parsed, directly progress full word section judgement, it is desirable that with the query statement that caches in ID mapping table is complete could unanimously obtain pair
Answer ID result.It is arranged whether judgement finally hits simultaneously, no hit then carries out conventional query and addition caching.
5. buffer consistency module as described in claim 1, which is characterized in that the main policies and content of buffer consistency are such as
Under:
Consistency is carried out mainly according to Invalidation protocols technology, notified when database has change client into
Row updates caching to reach strong consistency, for having the data of change, such as update and delete operation in database, records it
ID carries out buffer update algorithm.For there are newly-increased data in database, similitude comparison will be carried out, similitude comparison is pair
In newly-increased data, each attribute is inquired in the label of same type, there is the data record ID containing same alike result, it is believed that be high general
Rate set of metadata of similar data carries out buffer update.
6. caching replacement module as described in claim 1, which is characterized in that the main policies of caching replacement and content are as follows:
Each has corresponding creation time when being buffered in addition, time and number are used recently, according to Time-to-
Live fields technology, and not all caching is all high frequency, for example is logged in once, browser records its Entered state possible one
A month even some months will not all carry out identical inquiry again.For sporadic query statement, should swap out in time to reach caching
It is optimal.Therefore formulating following algorithm is considered sporadic sentence for the sentence in 24 hours without 3 access times, does not delay
Necessity is deposited, low recurrence sentence, the preferential array that swaps out of deposit are considered for the sentence in one week lower than 7 access times.When slow
When depositing that quantity is excessive or memory will overflow, inefficient caching removing algorithm is executed, the sentence preferentially to swap out in array is swapped out,
It will most remote be swapped out using the time recently according to according to LFU strategy simultaneously.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810981901.1A CN109446358A (en) | 2018-08-27 | 2018-08-27 | A kind of chart database accelerator and method based on ID caching technology |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810981901.1A CN109446358A (en) | 2018-08-27 | 2018-08-27 | A kind of chart database accelerator and method based on ID caching technology |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109446358A true CN109446358A (en) | 2019-03-08 |
Family
ID=65532875
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810981901.1A Pending CN109446358A (en) | 2018-08-27 | 2018-08-27 | A kind of chart database accelerator and method based on ID caching technology |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109446358A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110175206A (en) * | 2019-05-24 | 2019-08-27 | 江西尚通科技发展股份有限公司 | Intellectual analysis operational approach, system and medium for multiple database separation |
| CN110716940A (en) * | 2019-10-18 | 2020-01-21 | 成都九宽科技有限公司 | Incremental data access system |
| CN111831716A (en) * | 2020-06-17 | 2020-10-27 | 西安电子科技大学 | Data caching and retrieving method, system, medium, computer equipment and terminal |
| CN112788702A (en) * | 2021-01-07 | 2021-05-11 | 西安电子科技大学 | Mapping table item cache updating method for satellite Internet of things mobile terminal |
| CN113220710A (en) * | 2021-05-11 | 2021-08-06 | 北京百度网讯科技有限公司 | Data query method and device, electronic equipment and storage medium |
| WO2023028780A1 (en) * | 2021-08-30 | 2023-03-09 | Siemens Aktiengesellschaft | Method, apparatus and system for graph data caching |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102089741A (en) * | 2007-09-17 | 2011-06-08 | 国际商业机器公司 | Executing computer-intensive database user-defined programs on an attached high-performance parallel computer |
| US20150278396A1 (en) * | 2014-03-27 | 2015-10-01 | Elena Vasilyeva | Processing Diff-Queries on Property Graphs |
| CN105630881A (en) * | 2015-12-18 | 2016-06-01 | 陕西师范大学 | Data storage method and query method for RDF (Resource Description Framework) |
| CN106462535A (en) * | 2014-01-14 | 2017-02-22 | 口袋医生公司 | System and method for dynamic transactional data streaming |
| CN106599199A (en) * | 2016-12-14 | 2017-04-26 | 国云科技股份有限公司 | Data caching and synchronization method |
| CN106682343A (en) * | 2016-08-31 | 2017-05-17 | 电子科技大学 | Method for formally verifying adjacent matrixes on basis of diagrams |
| CN107194533A (en) * | 2017-04-07 | 2017-09-22 | 中国电力科学研究院 | A kind of power distribution network full information model building method and system |
| US20180068173A1 (en) * | 2016-09-02 | 2018-03-08 | VeriHelp, Inc. | Identity verification via validated facial recognition and graph database |
-
2018
- 2018-08-27 CN CN201810981901.1A patent/CN109446358A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102089741A (en) * | 2007-09-17 | 2011-06-08 | 国际商业机器公司 | Executing computer-intensive database user-defined programs on an attached high-performance parallel computer |
| CN106462535A (en) * | 2014-01-14 | 2017-02-22 | 口袋医生公司 | System and method for dynamic transactional data streaming |
| US20150278396A1 (en) * | 2014-03-27 | 2015-10-01 | Elena Vasilyeva | Processing Diff-Queries on Property Graphs |
| CN105630881A (en) * | 2015-12-18 | 2016-06-01 | 陕西师范大学 | Data storage method and query method for RDF (Resource Description Framework) |
| CN106682343A (en) * | 2016-08-31 | 2017-05-17 | 电子科技大学 | Method for formally verifying adjacent matrixes on basis of diagrams |
| US20180068173A1 (en) * | 2016-09-02 | 2018-03-08 | VeriHelp, Inc. | Identity verification via validated facial recognition and graph database |
| CN106599199A (en) * | 2016-12-14 | 2017-04-26 | 国云科技股份有限公司 | Data caching and synchronization method |
| CN107194533A (en) * | 2017-04-07 | 2017-09-22 | 中国电力科学研究院 | A kind of power distribution network full information model building method and system |
Non-Patent Citations (3)
| Title |
|---|
| ARTEM LYSENKO等: "representing and querying disease networks using graph databases", 《BIODATA MINING》 * |
| 金嘉辉: "面向大规模知识图谱的分布式查询技术研究", 《中国博士学位论文全文数据库 信息科技辑》 * |
| 黄权隆等: "HybriG:一种高效处理大量重边的属性图存储架构", 《计算机学报》 * |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110175206A (en) * | 2019-05-24 | 2019-08-27 | 江西尚通科技发展股份有限公司 | Intellectual analysis operational approach, system and medium for multiple database separation |
| CN110716940A (en) * | 2019-10-18 | 2020-01-21 | 成都九宽科技有限公司 | Incremental data access system |
| CN111831716A (en) * | 2020-06-17 | 2020-10-27 | 西安电子科技大学 | Data caching and retrieving method, system, medium, computer equipment and terminal |
| CN111831716B (en) * | 2020-06-17 | 2023-04-04 | 西安电子科技大学 | Data caching and retrieving method, system, medium, computer equipment and terminal |
| CN112788702A (en) * | 2021-01-07 | 2021-05-11 | 西安电子科技大学 | Mapping table item cache updating method for satellite Internet of things mobile terminal |
| CN112788702B (en) * | 2021-01-07 | 2022-07-22 | 西安电子科技大学 | Mapping table item cache updating method for satellite Internet of things mobile terminal |
| CN113220710A (en) * | 2021-05-11 | 2021-08-06 | 北京百度网讯科技有限公司 | Data query method and device, electronic equipment and storage medium |
| CN113220710B (en) * | 2021-05-11 | 2024-06-04 | 北京百度网讯科技有限公司 | Data query method, device, electronic equipment and storage medium |
| WO2023028780A1 (en) * | 2021-08-30 | 2023-03-09 | Siemens Aktiengesellschaft | Method, apparatus and system for graph data caching |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109446358A (en) | A kind of chart database accelerator and method based on ID caching technology | |
| CN104866434B (en) | Towards data-storage system and data storage, the call method applied more | |
| US10812589B2 (en) | Storage architecture for heterogeneous multimedia data | |
| US10666730B2 (en) | Storage architecture for heterogeneous multimedia data | |
| CN104750681B (en) | A kind of processing method and processing device of mass data | |
| CN102521334B (en) | Data storage and query method based on classification characteristics and balanced binary tree | |
| CN103177055B (en) | It is stored as row storage and row stores the hybrid database table of the two | |
| CN104850572B (en) | HBase non-primary key index construct and querying method and its system | |
| CN103177056B (en) | It is stored as row storage and row stores the hybrid database table of the two | |
| CN1552032B (en) | Database | |
| CN110162528A (en) | Magnanimity big data search method and system | |
| CN105868411A (en) | Non-relation type database and relation type database integrated data query method and system | |
| CN110362572A (en) | A kind of time series database system based on column storage | |
| CN109871367A (en) | A kind of distributed cold and heat data separation method based on Redis and HBase | |
| CN100594497C (en) | System for implementing network search caching and search method | |
| CN109947796B (en) | Caching method for query intermediate result set of distributed database system | |
| CN108664516A (en) | Enquiring and optimizing method and relevant apparatus | |
| CN107291964B (en) | A method of fuzzy query is realized based on HBase | |
| US20100274795A1 (en) | Method and system for implementing a composite database | |
| CN107656951B (en) | A kind of method of real time data in synchronous and heterogeneous Database Systems | |
| CN105138676B (en) | Table merge querying methods are divided in point storehouse for concurrently polymerizeing calculating based on high-level language | |
| CN110427437A (en) | A kind of relevant database mixing isomery interrogation model and method towards big data | |
| CN106599040A (en) | Layered indexing method and search method for cloud storage | |
| US11507555B2 (en) | Multi-layered key-value storage | |
| CN108763573A (en) | A kind of OLAP engines method for routing and system based on machine learning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190308 |
|
| WD01 | Invention patent application deemed withdrawn after publication |