CN102521303B - A kind of single-table multi-column sequence storage method for a column database - Google Patents
A kind of single-table multi-column sequence storage method for a column database Download PDFInfo
- Publication number
- CN102521303B CN102521303B CN201110392033.1A CN201110392033A CN102521303B CN 102521303 B CN102521303 B CN 102521303B CN 201110392033 A CN201110392033 A CN 201110392033A CN 102521303 B CN102521303 B CN 102521303B
- Authority
- CN
- China
- Prior art keywords
- row
- column
- data
- join index
- database
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 41
- 230000009467 reduction Effects 0.000 claims description 5
- 239000012141 concentrate Substances 0.000 claims description 2
- 230000008569 process Effects 0.000 description 8
- 238000005516 engineering process Methods 0.000 description 6
- 238000010586 diagram Methods 0.000 description 5
- 238000004458 analytical method Methods 0.000 description 4
- 239000000203 mixture Substances 0.000 description 3
- 238000011017 operating method Methods 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 238000007906 compression Methods 0.000 description 2
- 230000006835 compression Effects 0.000 description 2
- 238000013499 data model Methods 0.000 description 2
- 230000004899 motility Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 241001269238 Data Species 0.000 description 1
- 238000012356 Product development Methods 0.000 description 1
- 241001080526 Vertica Species 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000013144 data compression Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000008520 organization Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a kind of single-table multi-column sequence storage method for a column database.This column database includes multiple tables of data being made up of row and column, tables of data is divided into multiple row set, row set includes one or more row, and row union of sets collection constitutes tables of data, join index is set up in row set between any two, join index have recorded in two the row set establishing join index, belongs to the one-to-one relationship of the storage position of the row of the same tuple of tables of data.Utilize single table multiple row sequence provided by the present invention to store method, the search efficiency of column database can be improved, and reduce memory space.
Description
Technical field
The present invention relates to a kind of database storage method, particularly relate to a kind of for column database
Single table multiple row sequence storage method, belongs to database storage techniques field.
Background technology
Relevant database be one in order to the software system storing and processing structural data, its
Data are divided into two levels: one is logical data, and it is made up of tables of data, record etc.；Separately
One is physical data, and how its representation database stores logical data.Realize database physical number
According to method have two kinds: one be based on row storage, another kind be based on row storage.
For implementation method based on row storage, it is that the whole piece record of logical data is stored
In data block, in order to improve inquiry velocity, the index of the types such as B+ tree to be set up for some row；
For implementation method based on row storage, the record in logical data is the most directly mapped to thing by bar
In reason data, but record by row separately, the value of the same column of all records is existed together,
There is provided simultaneously and connect data, for the corresponding train value of difference record is reconfigured formation
Record.
Along with deepening continuously of enterprise and e-government, the complexity of database application increases day by day
By force.These demand drivings database technology and are developed to the direction of magnanimity and intelligence.Meanwhile, number
The data processing technique in the urgent need to real-time high-efficiency is applied according to warehouse and on-line analysis etc..Traditional
Database technology based on row storage has occurred in that technical bottleneck.How quickly to perform complexity
While inquiry, moreover it is possible to reduce memory space and cost-effective be current database technology research
Hot issue.
Column database is based on row memory technology, the relation that is mainly directed towards business decision analysis field
Type data base.The feature of row memory technology is that efficiency data query is high, reads disk few, storage sky
Between few, be the Ideal building data warehouse.It is right that the using value of column database comes from it
The storage advantage that quickly response and data compression are brought of complex query so that it is determine in enterprise
The applications such as plan analysis, data warehouse, business intelligence have good development prospect.According to
The analysis report about data warehouse that Gartner company of the U.S. issued in January, 2010: row
Data base, compared with traditional Relational DataBase, shows the performance of brilliance in terms of data analysis.
Therefore, technical research and the product development of column database is extensively closed at academia and industrial quarters
Note.
At present, the column database increased income has C-Store, rasdaman, MonetDB etc., business
Sybase IQ, Vertica Analytic Database, ParAccel is had by column database
Analytic Database, EXASOL EXA Solution etc..Over nearly 5 years, international one
The upper outstanding opinion about column database field of data base's meeting such as VLDB, SIGMOD, ICDE of stream
Literary composition occurs the most again and again.
In the Chinese invention patent application of Application No. 200810187227.6, disclose one
Realize the method and device of relevant database based on row storage, including: set up data file,
And the data block of composition data file is compiled serial number in order；Definition table segment；Record is inserted
In table segment；Record for being inserted in table segment is unique record identification number in generating table segment,
And by record by row separately；For each row in record, perform following operation: by train value
Store in data block with record identification number as Value Data and sort by train value size；By record
The serial number of the data block of identification number and storage Value Data stores new data as connecting data
In block, and press the sequence of record identification size；Data block and storage to storage Value Data connect
The data block of data sets up index, generates index data block.The method is to storage Value Data
Data block and storage connect the data block of data and set up index rather than to belonging to same tuple
Index is set up between different lines or row set.
Summary of the invention
The technical problem to be solved is to provide a kind of single table for column database many
Row sequence storage method.Utilize this storage method can improve the search efficiency of column database, and subtract
Little memory space.
For realizing above-mentioned goal of the invention, the present invention uses following technical scheme:
A kind of single-table multi-column sequence storage method for a column database, described column database includes many
The individual tables of data being made up of row and column, described tables of data is divided into multiple row set, described row
Set includes one or more row, and described row union of sets collection constitutes described tables of data, its
It is characterised by:
Join index is set up in described row set between any two, and described join index record establishes even
Connect in two row set of index, belong to the storage position of the row of the same tuple of described tables of data
One-to-one relationship.
The most more preferably, to each row set, according to belong to described tables of data same tuple,
The storage positional value of the row in two row set, sets up join index；
Described join index correspondence is stored described each row set in, and with described each row
Described row correspondence in set.
The most more preferably, if two row set comprise repeats row, then according to the described row that repeat, right
The row of said two row set is ranked up and stores；
If two row set do not repeat row, then by the row of said two row set, press respectively
Sort according to querying condition and store.
The most more preferably, for not repeating the said two row set of row, if logical order phase
Same then do not set up join index, if logical order differs, set up join index.
The most more preferably, for there being the said two row set repeating row, join index is not set up.
The most more preferably, estimate the cost needed for each implement plan, select even according to optimal cost
Connect index.
The most more preferably, the plurality of row set is appeared at the whole row judging described tables of data
And when concentrate, create Materialized View for each row set, complete the establishment of row set.
The most more preferably, during setting up join index, also include that following row set loads step
Rapid:
Step 1: at described tables of data loading data；
Step 2: materialization all arranges set Materialized View, including the storage position of materialization each row set；
Step 3: set up join index；
Step 4: delete the data of described tables of data；
Step 5: delete unwanted storage positional value.
The present invention has broken row storage to be needed to keep belonging to the train value of same logic tuple in each column
The identical restriction in position so that it is flexible that this list table multiple row sequence storage method adds in use
Property.The present invention can divide the projection of optimal sequence to strengthen according to the inquiry of form class application
Performance, also can process the inquiry of Ad-Hoc (point-to-point) class and not loss property according to join index
Energy.
Accompanying drawing explanation
The present invention is described in further detail with detailed description of the invention below in conjunction with the accompanying drawings.
Fig. 1 shows that the database items purpose of application this list table multiple row sequence storage method is shown
Example；
Fig. 2 is in the database project shown in Fig. 1, the schematic diagram of display storage location value；
Fig. 3 is in the database project shown in Fig. 1, shows the schematic diagram of join index；
Fig. 4 is in this list table multiple row sequence storage method, creates the operating procedure schematic diagram of row set；
Fig. 5 is in the embodiment shown in fig. 1, the schematic diagram of the join index set up；
Fig. 6 is in the embodiment shown in fig. 1, uses showing of join index reduction logic tuple
It is intended to.
Detailed description of the invention
The present invention uses the logic data model of relational database: each relation (relation) is
One bivariate table (table), by row (tuple, also referred to as tuple) and row, (attribute is also referred to as
Field) constitute.On this basis, the present invention uses physical organization based on row set to realize patrolling
Collect data model.Introduce the concrete meaning of " row set " first below.
Row set: each row set broadly falls into a relation, the above-listed set of logic is to close belonging to it
One vertical subset of system；Physically comprise one or more row and the affiliated pass of this relation
There is identical line number in system.If relation and another relation belonging to row set are many-to-one passes
System, can also comprise the row in another relation in row set.
Can be with duplicate packages containing same string between row set, it is possible to mutually do not repeat row.In other words
Saying, the row between row set can be overlapping, can comprise the same column of multiple tables of data.Belong to
Set in homonymous all row set, that the union of row is exactly this pass series, these row
Just constitute this relation.Row set uses row storage, it is possible to according in row set
Row or several row are ranked up.Such organizational form can be saved the storage overhead of index and carry
For the optimization space for inquiry.In storing process, it is possible to use multiple row storage compression side
Formula, such as RLE (run length encoding) etc..Further, it is possible to use the mode of fragmented storage
Improve compression efficiency.
The present invention provides a kind of row for OLAP (on-line analytical processing) scene to store physical set
Knit mode, less memory space offer can be used more to optimize motility.To this end,
In single table multiple row sequence storage method provided by the present invention, the tables of data conduct that first will store
Base table, is divided into the set of multiple row.Use storage positional value (storage key) simultaneously
Connection establishment join index (join index).This join index be used for recording different lines set it
Between the one-to-one relationship of train value.Row set is permissible by the join index connecting other row set
Obtain other train values that in this row set, train value is corresponding in base table, in order to rebuild in logic
Bar tuple.Storage positional value can set according to actual needs, such as, can be that base table is Sino-Kazakhstan
The cryptographic Hash etc. of uncommon row.
This row set is a vertical division of base table, comprises one or more row, and base table has
Identical line number.It addition, row set can also comprise other numbers having many-one relationship with base table
Row according to table.
Need to keep belonging to the train value of same logic tuple often owing to the present invention has broken row storage
Position in row needs identical restriction so that this list table multiple row sequence storage method can increase to be made
The motility used.The present invention can divide optimal sequence according to the inquiry of form class application
Projection, to strengthen performance, the most in figure 6, first carries out the projection of name, further according to connection
Index finds the place system of correspondence.The premise realizing this process is: record in join index
It is that row set place ties up to arrange the storage position in set name.Therefore, in each row set
Will store join index, the row in minute book row set arranges the position in gathering at another,
During i.e. the row in the set of these row is gathered with another row which is corresponding, also can be according to connecting strand
Draw and process the inquiry of Ad-Hoc (point-to-point) class and do not lose performance.
Single table multiple row sequence storage method of the present invention, sets up join index between any two in row set,
Join index have recorded, and establishes in two row set of join index, belongs to the same of tables of data
The one-to-one relationship of the storage position of the row of one tuple.To each row set, several according to belonging to
According to table same tuple, the storage positional value of row in two row set, set up join index,
Join index correspondence stored in each row set, and with the respective column in each row set
Value correspondence.Such that make each row set be resequenced, the value of join index is therewith
And sort, also would not affect the corresponding relation between the row belonging to same tuple of row set.
Below, as a example by the database project of application single table multiple row sequence storage method, specifically
Illustrate enforcement step and the effect thereof of the present invention.This database example is to impart knowledge to students for institution of higher learning
The data base of management work.As it is shown in figure 1, press student number sequence row set 1 (student number, name,
Sex, the age) and constitute student pass by the row set 2 (student number, place is) of place system sequence
System, contains the row in course in row set 3.As in figure 2 it is shown, each column arranged in set
Each value has a storage positional value, and the row with same storage positional value constitute a logic
Tuple.This storage positional value can be not use physical store mode.As it is shown on figure 3, due to
The row set of component relationship can have same storage position according to different row sequences in base table
Put the train value of value, different positions may be in different row set.Such as student number in Fig. 3
It is the record of 20070026, in row set 2, is positioned at the 1st row；It is positioned at the in row set 1
2 row.Use this position relationship of join index labelling, physical store position can not can determine that
When putting, find out the train value identical with storing positional value, and then build by closing at different row collection
Vertical join index, can obtain logic tuple according to certain order arranged.Such as, join index is (i.e.
In Fig. 3 top-right " correspondence position ") marked, in row set 2 in the 1st row and column set
The 2nd row corresponding.
By above explanation, can need not between two row set that logical order is identical
Join index.If two row set comprise repeats row, then repeat row according to this, to two row
The row of set is ranked up and stores；If two row set do not repeat row, then by two row
The row of set, sorts according to querying condition respectively and stores.Do not repeat two row set of row,
If logical order is identical, do not set up join index, if logical order differs, set up connection
Index.There are two the row set repeating row, because its sequence is identical, then need not join index.
Fig. 4 shows in this list table multiple row sequence storage method, creates the operating procedure of row set.
First, judge whether list is empty according to row set definition.If list is sky, return, as
Really list does not then take row set for sky, determines whether that whether the row of base table part are at base table row collection
Close in A, if result is for being, by row and fall in lines and gather in B, if result is for otherwise to feed back
Error message.It follows that whether the row checked further in each row set not quoting other tables
Belong to base table？Whether the row in each row set quoting other tables belong to reference list, reference list and
Whether base table has main foreign key relationship？Comprise whether row belong to list X？Whether there is other row set
Contain the row (whole row of base table appear in row set) of base table？In above inspection
The result looked into is in the case of being, by the row comprised and fall in lines set B in.If there is
The situation that the result checked is no, the most according to circumstances feedback error information.At row set B and base
In the case of the row set of table is identical, creates a Materialized View for each row set, take the next one
Row set repeats above-mentioned operating procedure.
It is shown below and creates some SQL statement using the base table of row set storage mode to be used
Example:
Statement illustrates:
CREATE TABLE table_name specifies the base table that table name is table_name.
It is row storage that WITH (ORIENTATION=COLUMN) specifies the storage mode of base table.
PROJECTIONS clause is used for creating row set, it is intended that include row name set, in set
The row comprised, for the row of sequence.
In the case of only using the row set of row of base table, use following query statement:
SELECT column_name...FROM table_name ORDER BY column_name；
Using in the case of the row set of the row of other tables, using following query statement:
SELECT column_name...FROM table_name JOIN other_table USING
(main foreign key column) ORDER BY column_name.
Fig. 5 shows in the embodiment shown in fig. 1, the schematic diagram of the join index set up.
Set up the SQL statement that join index used as follows:
CREATE JOIN INDEX index_name FROM projection_a TO projection_b；
This SQL statement creates row set projection_a to row set projection_b's
Join index index_name.
During setting up join index, the data load process of row set is such that
1) other table loading datas referred at base table and row set；
2) the row set Materialized View that materialization is whole, including the storage position of materialization each row set；
3) join index is set up；
4) delete base table and arrange the data of other tables that set refers to；
5) unwanted storage position is deleted.
The process selecting join index can i.e. estimate each holding to use rule-based optimizing mode
Cost needed for row plan, the resource spent by each implement plan is quantified by this cost, root
The join index of optimum is selected according to this cost.Owing to using join index can cause random data
Handling up (IO) in storehouse, therefore should use join index less.
Fig. 6 shows in the embodiment shown in fig. 1, uses join index reduction logic tuple
Operating process.This operating process comprises the following steps that
Step 10: find the row set comprising all target column required for inquiry, if it has,
This row set is then utilized to restore tuple；If it is not, enter next step；
In figure 6 particularly as follows: find and comprise inquiry all targets, i.e. " name ", " place system ",
Row set.
Step 11: what searching row sequence was identical comprises multiple row of all target column required for inquiry
Set；
Step 12: in the multiple row set obtained in a step 11, select to use join index
Minimum row collection charge-coupled (the row set 1 in such as Fig. 6 and row set 2)；
Step 13: after completing the projection of target column, utilizes join index reduction tuple.
By postponing the use of join index as far as possible, until after completing the projection of target column, can
To strengthen query performance.
Above single-table multi-column sequence storage method for a column database provided by the present invention is carried out
Detailed description.To those skilled in the art, without departing substantially from true spirit
On the premise of any obvious change that it is done, all by composition to patent right of the present invention
Infringement, corresponding legal responsibility will be undertaken.
Claims (8)
1. a single-table multi-column sequence storage method for a column database, described column database is for closing
It is type data base, including multiple tables of data being made up of row and column, it is characterised in that include following
Step:
The tables of data that will store, as base table, is divided into multiple row set, described row collection cooperation
For a vertical division of described base table, including one or more row；
Using the connection establishment join index of storage positional value, described join index is for record not
Between same column set, the row in the one-to-one relationship of train value and the set of these row is at another row collection
Position in conjunction；
To each row set, according to belong to described tables of data same tuple, two row set
In the storage positional value of row, set up join index；
Storing join index in each row set, described row set is by connecting other row set
Join index obtain other train values that train value in this row set is corresponding in base table；Wherein, will
Described join index correspondence store described each row set in, and with described each row set in
Described row correspondence.
2. single-table multi-column sequence storage method for a column database as claimed in claim 1, its
It is characterised by comprising the steps:
If two row set comprise repeats row, then repeat row according to described, said two is arranged
The row of set is ranked up and stores；
If two row set do not repeat row, then by the row of said two row set, press respectively
Sort according to querying condition and store.
3. single-table multi-column sequence storage method for a column database as claimed in claim 2, its
It is characterised by:
For not repeating the said two row set of row, if logical order is identical, do not set up
Join index, if logical order differs, sets up join index.
4. single-table multi-column sequence storage method for a column database as claimed in claim 2, its
It is characterised by:
For there being the said two row set repeating row, do not set up join index.
5. single-table multi-column sequence storage method for a column database as claimed in claim 1, it is special
Levy and be to include the following step creating row set:
Judge described tables of data whole row appear at the plurality of row union of sets concentrate time
Wait, create Materialized View for each row set, complete the establishment of row set.
6. single-table multi-column sequence storage method for a column database as claimed in claim 1, it is special
Levy and be during setting up join index also to include following row set load step:
Step 1: at described tables of data loading data；
Step 2: materialization all arranges set Materialized View, including the storage position of materialization each row set；
Step 3: set up join index；
Step 4: delete the data of described tables of data；
Step 5: delete unwanted storage positional value.
7. single-table multi-column sequence storage method for a column database as claimed in claim 1, it is special
Levy and be also to include the following step utilizing join index reduction logic tuple:
Step 1: find the row set comprising all target column required for inquiry, if it has, then profit
Tuple is restored with this row set；If it is not, enter next step；
Step 2: the multiple row comprising all target column required for inquiry finding row sequence identical collect
Close；
Step 3: in the multiple row set obtained in step 2, select to use join index minimum
Row collection is charge-coupled；
Step 4: after completing the projection of target column, utilizes join index reduction tuple.
8. single-table multi-column sequence storage method for a column database as claimed in claim 1, it is special
Levy and be:
Belong to the row of the same tuple of described tables of data, refer to the multiple row in row set.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201110392033.1A CN102521303B (en) | 2011-11-30 | 2011-11-30 | A kind of single-table multi-column sequence storage method for a column database |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201110392033.1A CN102521303B (en) | 2011-11-30 | 2011-11-30 | A kind of single-table multi-column sequence storage method for a column database |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102521303A CN102521303A (en) | 2012-06-27 |
| CN102521303B true CN102521303B (en) | 2016-08-10 |
Family
ID=46292224
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201110392033.1A Active CN102521303B (en) | 2011-11-30 | 2011-11-30 | A kind of single-table multi-column sequence storage method for a column database |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102521303B (en) |
Families Citing this family (21)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103218691B (en) * | 2013-04-26 | 2016-09-07 | 吉林市赢科信息技术有限责任公司 | A kind of embedded business intelligence information management system and management method |
| CN103425781B (en) * | 2013-08-20 | 2018-04-06 | 北京国双科技有限公司 | The data processing method and device of column position in software interface form |
| JP6287441B2 (en) * | 2014-03-26 | 2018-03-07 | 日本電気株式会社 | Database device |
| CN104021145B (en) | 2014-05-16 | 2018-03-06 | 华为技术有限公司 | The method and apparatus that a kind of mixed service concurrently accesses |
| KR102177190B1 (en) | 2014-06-30 | 2020-11-10 | 마이크로소프트 테크놀로지 라이센싱, 엘엘씨 | Managing data with flexible schema |
| CN104090954B (en) * | 2014-07-04 | 2019-02-05 | 用友网络科技股份有限公司 | The connection method of meter reading and the connection system of meter reading |
| CN105446991B (en) * | 2014-07-07 | 2018-10-30 | 阿里巴巴集团控股有限公司 | Date storage method, querying method and equipment |
| CN106326305A (en) * | 2015-06-30 | 2017-01-11 | 星环信息科技（上海）有限公司 | Storage method and equipment for data file and inquiry method and equipment for data file |
| CN105224649B (en) * | 2015-09-29 | 2019-03-26 | 北京奇艺世纪科技有限公司 | A kind of data processing method and device |
| CN105589965A (en) * | 2015-12-23 | 2016-05-18 | 中国地质大学（武汉） | Spatial data storage method based on column database |
| CN106933919B (en) * | 2015-12-31 | 2020-03-03 | 北京国双科技有限公司 | Data table connection method and device |
| CN106471501B (en) * | 2016-03-24 | 2020-04-14 | 华为技术有限公司 | Data query method, data object storage method and data system |
| CN106354829A (en) * | 2016-08-31 | 2017-01-25 | 天津南大通用数据技术股份有限公司 | Physic-chemical method and device of column storage database |
| CN108268586B (en) * | 2017-09-22 | 2020-06-16 | 阿里巴巴（中国）有限公司 | Data processing method, device, medium and computing equipment across multiple data tables |
| CN108009254A (en) | 2017-12-05 | 2018-05-08 | 北京百度网讯科技有限公司 | More indexing means and device, cloud system and computer-readable recording medium |
| CN110019218B (en) * | 2017-12-08 | 2023-08-25 | 阿里巴巴集团控股有限公司 | Data storage and query method and equipment |
| CN110413624A (en) * | 2019-08-07 | 2019-11-05 | 南京录信软件技术有限公司 | A method of the multiple row stored in association deposited based on column |
| CN111737290A (en) * | 2020-06-03 | 2020-10-02 | 北京人大金仓信息技术股份有限公司 | Query method and device for distributed database |
| CN112182028B (en) * | 2020-09-29 | 2024-07-05 | 北京人大金仓信息技术股份有限公司 | Data line number query method and device based on table of distributed database |
| CN113641681B (en) * | 2021-10-13 | 2022-02-22 | 南京大数据集团有限公司 | Space self-adaptive mass data query method |
| CN115827653B (en) * | 2022-11-25 | 2023-09-05 | 深圳计算科学研究院 | Pure column type updating method and device for HTAP and mass data |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101751406A (en) * | 2008-12-18 | 2010-06-23 | 赵伟 | Method and device for realizing column storage based relational database |
-
2011
- 2011-11-30 CN CN201110392033.1A patent/CN102521303B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101751406A (en) * | 2008-12-18 | 2010-06-23 | 赵伟 | Method and device for realizing column storage based relational database |
Non-Patent Citations (1)
| Title |
|---|
| hbase介绍;bfs198;《百度文库-专业资料-IT/计算机》;20110304;正文第一、二节 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102521303A (en) | 2012-06-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102521303B (en) | A kind of single-table multi-column sequence storage method for a column database | |
| CN100562874C (en) | A kind of increment data capturing method and system | |
| CN102495906A (en) | Incremental data migration method capable of realizing breakpoint transmission | |
| CN101751406B (en) | Method and device for realizing column storage based relational database | |
| CN103823823B (en) | Denormalization policy selection method based on Frequent Itemsets Mining Algorithm | |
| CN107391306A (en) | A kind of isomeric data library backup file access pattern method | |
| US7685138B2 (en) | Virtual cursors for XML joins | |
| CN103902544A (en) | Data processing method and system | |
| CN102411600B (en) | Data quality automatic detection method based on implication rule | |
| CN105631003A (en) | Intelligent index establishing, inquiring and maintaining method supporting mass data classification and counting | |
| CN103970853A (en) | Method and device for optimizing search engine | |
| Shanoda et al. | JOMR: Multi-join optimizer technique to enhance map-reduce job | |
| CN102169491A (en) | Dynamic detection method for multi-data concentrated and repeated records | |
| CN101408882B (en) | Method and system for searching authorization document | |
| CN102521356A (en) | Regular expression matching equipment and method on basis of deterministic finite automaton | |
| CN113282607A (en) | Optimized query method for database | |
| Doulkeridis et al. | On saying" enough already!" in mapreduce | |
| Rao et al. | Efficient iceberg query evaluation using set representation | |
| CN106354829A (en) | Physic-chemical method and device of column storage database | |
| Kläbe et al. | Patched Multi-Key Partitioning for Robust Query Performance. | |
| Jin et al. | Simultaneous optimization of complex mining tasks with a knowledgeable cache | |
| CN112487015B (en) | Distributed RDF system based on incremental repartitioning and query optimization method thereof | |
| CN103942294A (en) | Query planning method for mass data retrieval in intelligent traffic field | |
| Sun et al. | Mining schemas in semistructured data using fuzzy decision trees | |
| Zhao | Quotient Cube and QC-Tree: Efficient Summarizations for Semantic OLAP |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CP03 | Change of name, title or address | ||
| CP03 | Change of name, title or address |
Address after: 100102 201, 2 / F, 101, No. 5 building, No. 7 Rongda Road, Chaoyang District, Beijing Patentee after: China Electronics Technology Group Jincang (Beijing) Technology Co.,Ltd. Country or region after: China Address before: Room 601, Building 4, No. 8 Shangdi West Road, Haidian District, Beijing Patentee before: BEIJING KINGBASE INFORMATION TECHNOLOGIES Inc. Country or region before: China |