CN102521304A - Hash based clustered table storage method - Google Patents
Hash based clustered table storage method Download PDFInfo
- Publication number
- CN102521304A CN102521304A CN2011103922746A CN201110392274A CN102521304A CN 102521304 A CN102521304 A CN 102521304A CN 2011103922746 A CN2011103922746 A CN 2011103922746A CN 201110392274 A CN201110392274 A CN 201110392274A CN 102521304 A CN102521304 A CN 102521304A
- Authority
- CN
- China
- Prior art keywords
- hash
- tuple
- clusters
- row
- storage means
- 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
Images
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a hash based clustered table storage method, which comprises the following steps: step 1, an empty data table page space is pre-initialized; step 2, one or more columns of a data table are specified as a hash column or columns; step 3, a hash value is calculated according to the value of the hash column of each tuple to be used as the storage position of the corresponding tuple; step 4, mapping a row pointer of the corresponding tuple on a page space according to the hash value; and step 5, inserting the corresponding tuple into the page space according to the row pointer. According to the invention, the traditional hash clustering technology is further improved, so that a database storage management mechanism that the index is jumped over and the tuple is directly reached can be realized, and thus a large amount of cache resources occupied by the index in the use process of a large-scale database system is avoided, and the use performance of the database system is improved.
Description
Technical field
The present invention relates to a kind of database storing method, relate in particular to a kind of needs of data base-oriented storage administration, the table storage means that clusters based on Hash (hash) belongs to the database storage techniques field.
Background technology
Database (Database) is to organize and deposit the data acquisition in the second-level storage according to certain data model.In database technology, can use two kinds of extension data of formal description: physical data is described and logical data is described.Physical data is described and is meant the storage mode of data on memory device, and physical data is the actual data that leave on the memory device, and these data are also referred to as physical record.Logical data is described and is meant the data mode that user or programmer are used to operate, and logical data is a kind of abstract concept, is reflection and record to the objective reality world, and these data also can be called logical record.Conversion between physical data and the logical data realizes through data base management system (DBMS).
In data base management system (DBMS), adopt field to come the minimum unit of information that to name of mark-up entity attribute.The set of field is called tuple.Concrete entity of an element group representation.In existing relevant database, often with the row in the tables of data as tuple, row as field.A tables of data is made up of row (tuple) and row (field), forms a two-dimentional relation table.Several tables of data, view reach the Database Systems that are associated of a relevant unification of compositions such as file.
In Database Systems, index is a kind of structure that the row or the value of multiple row in the tables of data are sorted, and makes the customizing messages of index of reference in can the fast access tables of data.Index is divided into two kinds of clustered index and non-clustered index.What is called clusters and is meant in order to improve the inquiry velocity of certain field (or field groups), with leaving in the continuous physical block in the tuple set that has equal values on these fields.Therefore, clustered index can improve the speed of multirow retrieval, but not clustered index is suitable for the retrieval of single file.
Hash cluster (hash cluster) be meant through allocating the mode in space in advance, with the deposit data of same keyword (key) together, to improve a technology of query performance.At present, only product has the Hash function that clusters in Oracle series data storehouse, other database product, for example SQL Server, IBM DB2 and reach among dream DM etc. and all do not have similar functions.In actual use, still there is certain defective in this technology, and for example the quantity of key word (key) is difficult to accurate estimation, causes the application scenarios of Hash clustering strategies very limited.
Summary of the invention
In view of existing in prior technology is not enough, technical matters to be solved by this invention is to provide a kind of table storage means that clusters based on Hash.This method can provide the database storing administrative mechanism that strides across the through tuple of index.
For realizing above-mentioned goal of the invention, the present invention adopts following technical scheme:
A kind of table storage means that clusters based on Hash, said tables of data is made up of tuple and row, it is characterized in that may further comprise the steps:
Step 1: the tables of data page space that initialization in advance is empty;
Step 2: one or more Hash row of classifying as of specifying said tables of data;
Step 3: the value according to the said Hash row of each tuple is calculated cryptographic hash, as the memory location of respective tuple;
Step 4: shine upon the line pointer of said respective tuple on said page space according to said cryptographic hash;
Step 5:, said respective tuple is inserted into said page space according to said line pointer.
Wherein more excellently, according to the number of the cryptographic hash tuple that possibly use, the Hash row, said page space is carried out dynamic predistribution.
Wherein more excellently, in said step 3,, chain is overflowed in corresponding cryptographic hash increase, this tuple is stored into overflow in the chain when occurring two or more cryptographic hash in the cryptographic hash of the Hash of each tuple row when identical.
Wherein more excellently, also comprise the query steps that clusters and show said:
Value according to the Hash of appointment row calculates corresponding cryptographic hash, finds line pointer through said cryptographic hash according to the mapping relations of having set up, finds corresponding tuple according to said line pointer.
Wherein more excellently, when the said table that clusters is that non-unique types clusters when table, said query steps also comprises: the value of the Hash row of many tuples that inquiry obtained, meet line pointer is verified.
Wherein more excellently, when the said table that clusters is that unique types clusters when table, the value of the Hash row of many tuples that meet line pointer that inquiry is obtained is not verified.
Wherein more excellently, if the operation of inserting is arranged at the back, so also comprise the cluster step of page space of table of dynamic expansion at query script:
The value of the Hash row of the tuple that acquisition need to be inserted calculates cryptographic hash then, and obtains the said tuple that needs to insert through man-to-man mapping and be stored in the line pointer on the page, finds corresponding page to insert according to said line pointer then.
Wherein more excellently, if the page space of the page that said line pointer finds is not enough, then the said tuple that need insert is stored in and overflows in the chain.
Wherein more excellently, if inquiry row comprise all Hash row of said tables of data, prefix and sorting operation that the Hash row are classified in the inquiry when perhaps using specific hash function as use the scan mode that clusters to scan.
The table storage means that clusters provided by the present invention has been done further improvement to existing Hash clustering strategies; Can realize striding across the database storing administrative mechanism of the through tuple of index; Thereby in the use of large scale database system, avoided index to the taking in a large number of cache resources, improved the usability of Database Systems.
Description of drawings
Below in conjunction with accompanying drawing and embodiment the present invention is done further detailed description.
Fig. 1 obtains the synoptic diagram of tuple in the tables of data for the table that clusters through Hash.
Embodiment
In the use of large scale database system, index itself will take a large amount of cache resources when index of reference was scanned, thereby had reduced the utilization factor of cache resources.In order to address this problem, need to realize to stride across the database storing administrative mechanism of the through tuple of index.For this reason; The table storage means that clusters provided by the present invention is at first set up cryptographic hash (the harsh value through the tuple row; Cryptographic hash) to a kind of mapping of page stores position; The value of the inquiry row of appointment can directly calculate the position of tuple in the page when making according to inquiry, thereby reaches the purpose of direct location tuple.Bright specifically in the face of this expansion down.
In the table storage means that clusters provided by the present invention, said tables of data is made up of tuple and row, specifically may further comprise the steps:
Step 1: the tables of data page space that initialization in advance is empty;
Step 2: one or more Hash row of classifying as of specifying said tables of data;
Step 3: the value according to the said Hash row of each tuple is calculated cryptographic hash, as the memory location of respective tuple;
Step 4: shine upon the line pointer of said respective tuple on said page space according to said cryptographic hash;
Step 5:, said respective tuple is inserted into said page space according to said line pointer.
Particularly, at first be the empty tables of data page space of initialization in advance in foundation clusters the process of showing.Next is the one or more hash of classifying as the row of specific data table, according to the cryptographic hash of the hash row of each tuple, confirms the position that tuple is stored, so that tuple is inserted in the assigned address according to cryptographic hash.
For the use of the table of supporting to cluster, the present invention adopted table predistribution technology and table overflow the chain treatment technology.The predistribution technology of table is meant that when the establishment table promptly according to the dynamic predistribution of cryptographic hash space, tuple can directly be inserted into assigned address when operating like this, safeguarded the characteristic that tuple clusters according to the cryptographic hash of index column.
The chain treatment technology that overflows of table then is the special processing to the table that clusters.In particular cases, in the cryptographic hash of the hash row of each tuple that setting up clusters obtains when showing, it is identical to have two or more cryptographic hash, and this will make corresponding two or more tuples of certain cryptographic hash.When the corresponding tuple of certain cryptographic hash is too much, only need overflow chain to corresponding cryptographic hash increase, this tuple is stored into overflow in the chain, and need not have influence on the maintenance of the tuple of other cryptographic hash.
At the beginning of foundation clustered table, the cluster storage space of table of the dynamic predistribution of number of the cryptographic hash of the Hash row of the tuple that can possibly use according to database application made that so in most cases tuple can directly be located according to the cryptographic hash that hash is listed as.The situation that the cryptographic hash that is listed as for hash repeats if in the pre-assigned page, there are not enough space storage tuples, then can be carried out the storage of tuple in overflowing chain.But in this case,, possibly cause the performance of Database Systems to descend owing to can not realize direct location to tuple.
In the table storage means that clusters provided by the present invention, can use two types the Hash table that clusters: the Hash of non-unique (uniqueness) type clusters and shows and the Hash of unique (uniqueness) the type table that clusters.Two kinds of Hash cluster and show on storage mode, to exist certain difference.In the former, a plurality of tuples can corresponding cryptographic hash, therefore will reserve the space of a plurality of tuples in the same cryptographic hash groove, but also the support that possibly overflow chain, somewhat complicated in practical operation, but versatility is stronger.In the latter, different tuples must have the different Hash value.Because have the different restriction of cryptographic hash of different tuples, range of application is smaller, in case but application then can play the effect of the current tuple in direct location has greatly promoted the management of performance of Database Systems.The user can select the suitable table type that clusters according to practical application.
C_stock table in using with TPCC (benchmark test) below is an example, specifies the implementation procedure of the table storage means that originally clusters.Suppose this c_stock table have a major key (s_w_id, s_i_id), two row are integers, and in TPCC used, the value of these two row was followed successively by (1,1), (1; 2) ..., (1,100000), (2,1) ... (2,100000) ..., (n, 100000), 100000n bar tuple altogether.
Based on following technical reason, select the Hash of the unique type table that clusters to set up this c_stock table, and the hash row be (s_w_id, s_i_id):
(1) better regularity of the value of primary key column.The distribution of primary key column is very regular, can avoid the cryptographic hash of Hash row of different tuples identical.It adopts the sequence that increases progressively, the hash function that therefore can use integer hash row increasing function to distribute as tuple.Because the s_i_id row are that 100000 carries arrive s_w_id, so this integer hash row increasing function that generates automatically is f (s_w_id, s_i_id)=(s_w_id-1) * 100000+ (s_i_id-1).
(2) directive property of querying condition is good, promptly can obtain a corresponding tuple according to querying condition.In the query script that TPCC uses, be that querying condition obtains a corresponding tuple and inquires about or upgrade operation how with the values of two row, therefore can directly use the Hash scanning algorithm as a rule, reach the purpose of optimizing data throughput (IO).
(3) query script does not have the operation of insertion.In the query script that TPCC uses; Can on the c_stock table, not insert operation; Therefore the bar of c_stock table remains unchanged after counting and pouring into data before self-test begins; When building table, can calculate the actual size of c_stock table, in subsequent operation, need not continue to expand the usage space of this c_stock table.
In the process of setting up the c_stock table; The space of initialization 100000n bar tuple when creating table at first; The value that when pouring into data, is listed as according to hash is calculated the cryptographic hash (hash value) of hash function; Shine upon line pointer then so that carrying out tuple inserts, guarantee that all tuples are inserted in the c_stock table according to the order of hash train value.At this moment, needed data have promptly been filled up in the c_stock table.
As shown in Figure 1; After the c_stock of the table that clusters as Hash table is created completion, inquire about if desired, can calculate corresponding cryptographic hash according to the value of hash row; Find line pointer through the mapping relations of having set up then, and then find corresponding tuple according to line pointer.Because c_stock table is the Hash of the unique type table that clusters, and use integer hash row increasing function is as hash function, so needn't carry out the checking of s_w_id and s_i_id train value when finding respective tuple according to line pointer.
In the embodiment of table storage means that originally clusters, for following inquiry:
SELECT*from?c_stock?where?s_w_id＝5?order?by?s_i_id；
Then directly calculate cryptographic hash, and then carry out the location of tuple according to hash train value (5,1).Because promptly according to the rank order of s_i_id, sequencer procedure is promptly accomplished in the scanning of therefore directly carrying out on the heap to the order of the tuple in the heap (heap), needs the tuple of scanning to be 100000 at most.
In scanning process, originally cluster the table storage means except can using traditional sequential scanning, index scanning, can also use the scan mode that clusters.Comprise all hash row of table for inquiry row, when prefix and the sorting operation of hash row classified in inquiry as during perhaps specific hash function, the preferred scan mode that clusters of using was carried out, thereby directly locatees tuple.
In Hash clusters the query script of table,, then need verify its whether satisfied condition of inquiring about one by one to many tuples that meet line pointer if be the Hash of the non-unique type table that clusters.If, may scan several pages, the performance of Database Systems is caused certain influence overflowing under the situation that chain exists more.And, used integer hash row increasing function simultaneously as hash function for the Hash of the unique type table that clusters, and then show an only tuple of corresponding corresponding train value of a line pointer, needn't verify the satisfying property of its querying condition, can directly return.
Based on the inquiry of the equivalence of cryptographic hash, can navigate to corresponding tuple for some according to the value of the hash row that calculate is disposable.With respect to using the indexed data storehouse to use, when saving index stores and reading, reached the effect of direct location tuple when making index of reference like this.
If the operation of insertion is arranged at the back, then can dynamically expand the page space that Hash clusters and shows at query script.When carrying out tuple insertion operation; At first obtain the value of the corresponding hash row of tuple; The hash function of invoke user appointment calculates cryptographic hash then, and obtains tuple through man-to-man mapping and be stored in the line pointer on the page, finds corresponding page to insert according to line pointer then.If the current page space is not enough, then can introduce and overflow chain mechanism, tuple is stored in overflows in the chain.
When carrying out Hash and cluster the operations such as renewal, deletion of table, if in filtercondition, indicated the value of hash row, then equally can application class like the process of inquiry, find the page at tuple place to carry out corresponding operation.For example when upgrading operation, carry out following SQL statement:
UPDATE?c_stock?SET?s_quantity＝36?WHERE?s_i_id＝38426?AND?s_w_id＝3；
It is 338425 that the value (3,38426) that then is listed as according to hash calculates cryptographic hash, thereby calculates corresponding line pointer, can obtain corresponding tuple and upgrade operation.
In the present invention, can use dissimilar hash functions.Except the hash function of acquiescence, the present invention can also use user-defined hash function or integer hash row increasing function.The use of the two kinds of functions in back makes that the storage of tuple is visible basically concerning the user.Cluster for the table for the Hash of unique type, normally used is user-defined hash function.The user can select suitable hash function, through the Hash that the forms the unique type as far as possible table that clusters, thereby promotes database performance as much as possible.
More than to provided by the present invention based on Hash cluster the table storage means carried out detailed explanation.To those skilled in the art, any conspicuous change of under the prerequisite that does not deviate from connotation of the present invention, it being done all will constitute to infringement of patent right of the present invention, with corresponding legal responsibilities.
Claims (9)
1. table storage means that clusters based on Hash, said tables of data is made up of tuple and row, it is characterized in that may further comprise the steps:
Step 1: the tables of data page space that initialization in advance is empty;
Step 2: one or more Hash row of classifying as of specifying said tables of data;
Step 3: the value according to the said Hash row of each tuple is calculated cryptographic hash, as the memory location of respective tuple;
Step 4: shine upon the line pointer of said respective tuple on said page space according to said cryptographic hash;
Step 5:, said respective tuple is inserted into said page space according to said line pointer.
2. the table storage means that clusters as claimed in claim 1 is characterized in that:
According to the number of the cryptographic hash tuple that possibly use, the Hash row, said page space is carried out dynamic predistribution.
3. the table storage means that clusters as claimed in claim 1 is characterized in that:
In the said step 3,, chain is overflowed in corresponding cryptographic hash increase, this tuple is stored into overflow in the chain when occurring two or more cryptographic hash in the cryptographic hash of the Hash of each tuple row when identical.
4. the table storage means that clusters as claimed in claim 1 is characterized in that:
Also comprise query steps to the said table that clusters:
Value according to the Hash of appointment row calculates corresponding cryptographic hash, finds line pointer through said cryptographic hash according to the mapping relations of having set up, finds corresponding tuple according to said line pointer.
5. the table storage means that clusters as claimed in claim 4 is characterized in that:
When the said table that clusters is that non-unique types clusters when table, said query steps also comprises:
The value of the Hash row of many tuples that inquiry obtained, meet line pointer is verified.
6. the table storage means that clusters as claimed in claim 4 is characterized in that:
When the said table that clusters is that unique types clusters when table, the value of the Hash row of many tuples that meet line pointer that inquiry is obtained is not verified.
7. the table storage means that clusters as claimed in claim 4 is characterized in that:
If the operation of inserting is arranged, then also comprise the cluster step of page space of table of dynamic expansion behind query script:
The value of the Hash row of the tuple that acquisition need to be inserted calculates cryptographic hash then, and obtains the said tuple that needs to insert through man-to-man mapping and be stored in the line pointer on the page, finds corresponding page to insert according to said line pointer then.
8. the table storage means that clusters as claimed in claim 7 is characterized in that:
If the page space of the page that said line pointer finds is not enough, then the said tuple that need insert is stored in and overflows in the chain.
9. the table storage means that clusters as claimed in claim 1 is characterized in that:
If inquiry row comprise all Hash row of said tables of data, prefix and sorting operation that the Hash row are classified in the inquiry when perhaps using specific hash function as then use the scan mode that clusters to scan.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2011103922746A CN102521304A (en) | 2011-11-30 | 2011-11-30 | Hash based clustered table storage method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2011103922746A CN102521304A (en) | 2011-11-30 | 2011-11-30 | Hash based clustered table storage method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN102521304A true CN102521304A (en) | 2012-06-27 |
Family
ID=46292225
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2011103922746A Pending CN102521304A (en) | 2011-11-30 | 2011-11-30 | Hash based clustered table storage method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102521304A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104021161A (en) * | 2014-05-27 | 2014-09-03 | 华为技术有限公司 | Cluster storage method and device |
| CN104462124A (en) * | 2013-09-22 | 2015-03-25 | 中国电信股份有限公司 | Data storage platform organization method based on linear hash table and data storage platform |
| CN105335475A (en) * | 2015-09-30 | 2016-02-17 | 中国科学院计算技术研究所 | Method and system for locality non-cluster index based on streaming data |
| CN106302179A (en) * | 2016-07-29 | 2017-01-04 | 杭州迪普科技有限公司 | A kind of method and device managing concordance list |
| CN108304460A (en) * | 2017-12-25 | 2018-07-20 | 中国电力科学研究院有限公司 | A kind of localization method and system improving database |
| CN112256698A (en) * | 2020-10-16 | 2021-01-22 | 美林数据技术股份有限公司 | Automatic table relation association method based on multi-Hash function |
| CN114356226A (en) * | 2021-12-17 | 2022-04-15 | 广州文远知行科技有限公司 | Sensor data storage method, device, equipment and storage medium |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1926517A (en) * | 2004-03-12 | 2007-03-07 | 国际商业机器公司 | Method and system for affinity management |
| CN101477524A (en) * | 2008-12-11 | 2009-07-08 | 金蝶软件（中国）有限公司 | System performance optimization method and system based on materialized view |
-
2011
- 2011-11-30 CN CN2011103922746A patent/CN102521304A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1926517A (en) * | 2004-03-12 | 2007-03-07 | 国际商业机器公司 | Method and system for affinity management |
| CN101477524A (en) * | 2008-12-11 | 2009-07-08 | 金蝶软件（中国）有限公司 | System performance optimization method and system based on materialized view |
Non-Patent Citations (3)
| Title |
|---|
| EDWINGU: "《哈希聚簇读取（Hash Cluster Access）[摘]》", 《EDWINKOO的专栏》 * |
| JACKY: "《Oracle Cluster使用场景分析》", 《HELLO DATABASE》 * |
| SHARANAM SHAH ETAL,: "《Oracle for Professionals-Covers Oracle 9i，10g&11g》", 17 June 2008 * |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104462124A (en) * | 2013-09-22 | 2015-03-25 | 中国电信股份有限公司 | Data storage platform organization method based on linear hash table and data storage platform |
| CN104462124B (en) * | 2013-09-22 | 2018-04-06 | 中国电信股份有限公司 | Data storing platform method for organizing and data storing platform based on linear Hash table |
| CN104021161A (en) * | 2014-05-27 | 2014-09-03 | 华为技术有限公司 | Cluster storage method and device |
| WO2015180432A1 (en) * | 2014-05-27 | 2015-12-03 | 华为技术有限公司 | Clustering storage method and device |
| US10817258B2 (en) | 2014-05-27 | 2020-10-27 | Huawei Technologies Co., Ltd. | Clustering storage method and apparatus |
| RU2663358C2 (en) * | 2014-05-27 | 2018-08-03 | Хуавэй Текнолоджиз Ко., Лтд. | Clustering storage method and device |
| CN105335475B (en) * | 2015-09-30 | 2018-07-10 | 中国科学院计算技术研究所 | A kind of locality Nonclustered index method and system based on stream data |
| CN105335475A (en) * | 2015-09-30 | 2016-02-17 | 中国科学院计算技术研究所 | Method and system for locality non-cluster index based on streaming data |
| CN106302179A (en) * | 2016-07-29 | 2017-01-04 | 杭州迪普科技有限公司 | A kind of method and device managing concordance list |
| CN106302179B (en) * | 2016-07-29 | 2020-02-11 | 杭州迪普科技股份有限公司 | Method and device for managing index table |
| CN108304460A (en) * | 2017-12-25 | 2018-07-20 | 中国电力科学研究院有限公司 | A kind of localization method and system improving database |
| CN108304460B (en) * | 2017-12-25 | 2022-10-25 | 中国电力科学研究院有限公司 | Improved database positioning method and system |
| CN112256698A (en) * | 2020-10-16 | 2021-01-22 | 美林数据技术股份有限公司 | Automatic table relation association method based on multi-Hash function |
| CN112256698B (en) * | 2020-10-16 | 2023-09-05 | 美林数据技术股份有限公司 | Table relation automatic association method based on multi-hash function |
| CN114356226A (en) * | 2021-12-17 | 2022-04-15 | 广州文远知行科技有限公司 | Sensor data storage method, device, equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102129458B (en) | Method and device for storing relational database | |
| CN110321344B (en) | Information query method and device for associated data, computer equipment and storage medium | |
| CN102521304A (en) | Hash based clustered table storage method | |
| CN108874971B (en) | Tool and method applied to mass tagged entity data storage | |
| JP6025149B2 (en) | System and method for managing data | |
| CN107818115B (en) | Method and device for processing data table | |
| CN102968503B (en) | The data processing method of Database Systems and Database Systems | |
| EP2069979B1 (en) | Dynamic fragment mapping | |
| CN101840400B (en) | Multilevel classification retrieval method and system | |
| CN104112008A (en) | Multi-table data association inquiry optimizing method and device | |
| EP2444906A1 (en) | Mapping of table data to hierarchical format for graphical representation | |
| CN103177059A (en) | Split processing paths for database calculation engine | |
| CN104102710A (en) | Massive data query method | |
| CN102362273A (en) | Dynamic hash table for efficient data access in relational database system | |
| US9235613B2 (en) | Flexible partitioning of data | |
| CN103914483B (en) | File memory method, device and file reading, device | |
| CN1955958A (en) | Sort data storage and split catalog inquiry method based on catalog tree | |
| CN103440245A (en) | Line and column hybrid storage method of database system | |
| CN104063376A (en) | Multi-dimensional grouping operation method and system | |
| CN102737123B (en) | A kind of multidimensional data distribution method | |
| Wang et al. | Distributed storage and index of vector spatial data based on HBase | |
| CN104572785A (en) | Method and device for establishing index in distributed form | |
| CN113177090A (en) | Data processing method and device | |
| CN114185895A (en) | Data import and export method and device, electronic equipment and storage medium | |
| CN104573112A (en) | Page query method and data processing node for OLTP cluster database |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C12 | Rejection of a patent application after its publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20120627 |