CN109165222A - A kind of HBase secondary index creation method and system based on coprocessor - Google Patents
A kind of HBase secondary index creation method and system based on coprocessor Download PDFInfo
- Publication number
- CN109165222A CN109165222A CN201810945470.3A CN201810945470A CN109165222A CN 109165222 A CN109165222 A CN 109165222A CN 201810945470 A CN201810945470 A CN 201810945470A CN 109165222 A CN109165222 A CN 109165222A
- Authority
- CN
- China
- Prior art keywords
- index
- data
- region
- line unit
- coprocessor
- 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
- 238000000034 method Methods 0.000 title claims abstract description 12
- 238000003780 insertion Methods 0.000 claims abstract description 16
- 230000037431 insertion Effects 0.000 claims abstract description 16
- 238000000926 separation method Methods 0.000 claims abstract description 8
- 230000006870 function Effects 0.000 claims description 16
- 238000004364 calculation method Methods 0.000 claims description 6
- 238000005303 weighing Methods 0.000 claims description 2
- 238000013500 data storage Methods 0.000 description 3
- 230000000295 complement effect Effects 0.000 description 2
- 238000013523 data management Methods 0.000 description 2
- 230000000630 rising effect Effects 0.000 description 2
- 239000007787 solid Substances 0.000 description 2
- 230000001680 brushing effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 239000002360 explosive Substances 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
Index data and master data are carried out logical separation according to pre- subregion and random hash strategy by the present invention relates to a kind of HBase secondary index creation method and system based on coprocessor;According to the different column families in same table, index data and master data are subjected to physical separation.The system includes: insertion module, for constructing secondary index according to index configurations file, index data being inserted into secondary index area, newly-generated master data is inserted into main data area in data insertion;Further include: enquiry module, for constructing querying condition according to index configurations file in data query, after the secondary index area of parallel query Region obtains index line unit, the asynchronous acquisition master data on Region.A kind of HBase secondary index creation method and system based on coprocessor proposed by the present invention efficiently, rapidly can carry out field search to HBase.
Description
Technical field
The present invention relates to database technical field, especially a kind of HBase secondary index creation side based on coprocessor
Method and system.
Background technique
The arriving of big data era has pushed the multi-field theoretical high speed development with engineering practice such as data storage, processing,
Following data continue into explosive growth, and traditional data storage and management method has been difficult to adapt to current extensive
Demand of the data management to efficiency.For this purpose, non-relational NoSQL database is rapidly developed.HBase is as NoSQL data
The representative in library, has been widely used in the data storage and management of all trades and professions.Compared with traditional database, HBase can only
It is inquired according to the range of line unit or line unit, it is excessively single, it is not flexible, in most cases it is necessary to arrange HBase
Value inquiry, and HBase needs to use more advanced filter for arranging without establishing index to inquire.Filter needs pair
Full table is scanned, and search efficiency is lower, reduces the performance that HBase table is inquired by train value, and cause machine physical resource
Waste.
Currently, the scheme for establishing secondary index on HBase mainly has scheme based on third party's independent engine, based on association
The scheme of processor, the scheme intercepted based on memory and scheme based on complementary cluster formula etc..
Secondary index scheme based on third independent engine has ElasticSearch and Solr.ElasticSearch and
Solr is the full-text search server based on Lucene, by the field of inquiry involved in HBase table in ElasticSearch or
It is indexed in Solr.But this mode needs to safeguard a set of index cluster, causes overhead.
The Hindex that secondary index scheme based on coprocessor mainly has Huawei to propose.The program divides data and index
It opens and is stored in different tables, after being inserted into data in main table, write index column in another concordance list with coprocessor.But
It is that this sets of plan needs to modify HBase source code, invasive larger.Meanwhile in Region division, need to keep indexing
Region and the cut-off of data Region are logically consistent.
The IHBASE that the scheme intercepted based on memory mainly has YoramKulbak and DanWashusen to propose.The program exists
Region rank establishes index rather than table level is other, has been filled with when brushing into disk inside, will do it interception request, and in memory
Data construct index, index be stored in table in a manner of another column family.But need to reconstruct HBase, and several recently
Year does not all update.
Scheme based on complementary cluster formula mainly has the Computer Department of the Chinese Academy of Science to propose CCIndex.Detailed letter of the program data
Breath is also stored in concordance list, does not need to go in former table to go again by the line unit of acquisition to search data.But it is deleted more in data
When new, safeguard that the data in concordance list are more complicated.Simultaneously as its copy mechanism for having disabled bottom HDFS, causes data
Reliability decrease.
Summary of the invention
The purpose of the present invention is to provide a kind of HBase secondary index method and system based on coprocessor, with gram
Take defect existing in the prior art.
To achieve the above object, the technical scheme is that a kind of HBase secondary index creation based on coprocessor
Index data and master data are carried out logical separation according to pre- subregion and random hash strategy by method;According in same table not
Index data and master data are carried out physical separation by same column family.
In an embodiment of the present invention, two are logically divided on the same Region according to pre- subregion and random hash
Grade index area and main data area, secondary index area are used to store index data, and main data area is used to store master data.
In an embodiment of the present invention, same table is divided into two column families, a column family is used to store index data, separately
One column family is used to master data.
It further, further include a kind of HBase secondary index creation system based on coprocessor, comprising:
It is inserted into module, for secondary index being constructed according to index configurations file, index data being inserted into data insertion
To secondary index area, newly-generated master data is inserted into main data area;
Enquiry module, for constructing querying condition, parallel query Region according to index configurations file in data query
Secondary index area obtain index line unit after, the asynchronous acquisition master data on Region.
In an embodiment of the present invention, the insertion module is indexed insertion in accordance with the following steps:
Step 11: when building table, calculating the number of Region, pre- subregion is carried out to each Region;
Step 12: after pre- subregion, obtaining the section startKey and endKey of each Region;
Step 13: rewriteeing prePut () Hook Function of coprocessor, master data is inserted by function acquisition
rowkey;
Step 14: a value in the section startKey and endKey of Region is randomly generated, and is denoted as hashKey;
By hashKey splicing before being inserted into the rowkey of master data, and remember that the line unit being newly inserted into is hashRowkey;
Step 15: the search index configuration file in prePut () Hook Function, after obtaining the field indexed, parsing
Put object judges whether there is index field insertion;If so, the startKey in the section Region where hashKey is then obtained, and
The value of index field is spliced before being inserted into the line unit of master data, as index line unit;Otherwise directly put object is inserted into and is led
Data field;
Step 16: obtaining starting line unit, index train value and be inserted into the length value of line unit, splice according to following format: rising
Begin key, and _ index train value _ is inserted into line unit, and is inserted into a column of index column family;
Step 17: index data being inserted into index column family, master data is inserted into data column family;The title of index column family is solid
Fixed, do not allow to modify and do not allow name of weighing with index column Praenomen.
In an embodiment of the present invention, further include following steps in the step S11:
Step S111: determine that the pre- subregion number of cluster, the pre- subregion number calculation formula of individual node are as follows:
Wherein, M indicates the memory size of RegionServer;F indicates that RegionServer gives the ratio of memstore;S
Indicate the size of memstore, unit M;A is the number of column family in table;
Step S112: determine that the node number of cluster, the calculation formula of the total pre- subregion number of cluster are as follows:
R=P*N
Wherein, R indicates the total number of the pre- subregion of cluster, and P indicates the number of the pre- subregion of each node, and N indicates to save in cluster
The number of point.
In an embodiment of the present invention, further include following steps in the step S13:
Step S131: client issues put request;
Step S132: the request is assigned to corresponding RegionServer and Region;
Step S133: coprocessor intercepts the request, then on each RegionObserver online on the table
PrePut () Hook Function is called, put request is intercepted, parses put object, obtain rowkey value;
Step S134: if do not intercepted by prePut () Hook Function, put request continues to be sent to Region, then
It is handled.
In an embodiment of the present invention, the enquiry module, which includes the following steps, is inquired:
Step 21: when querying condition is arranged in client, querying condition being parsed by enquiring component, reads index configurations text
Part judges that the field whether there is in index configurations file;
Step 22: if it does, indicating that the field establishes index, then constructing querying condition by search index device
It originates line unit and terminates line unit;The startKey that the starting line unit of querying condition is each Region splices querying condition, looks into
The startKey that line unit is each Region that terminates of inquiry condition splices a value greater than querying condition, and goes to step
S24;
Step 23: if it does not exist, then carrying out full table scan, and going to step S27;
Step 24: after constructing new querying condition, the secondary index area in multi-thread concurrent to each Region is carried out
Scan inquiry;
Step 25: inquiring qualified index line unit in secondary index area, this is then parsed by the train value indexed
The line unit of the master data of record;
Step 26: asynchronous to obtain master data in batches on the Region after the index line unit for finding a Region
Record;
Step 27: result being pooled at enquiring component, then is pooled to client return.
Compared to the prior art, the invention has the following advantages:
1. the secondary index on data insertion speed based on coprocessor is better than being based on third-party secondary index;
2. for the secondary index based on coprocessor using concurrently inquiring, inquiry velocity is better than base in data query speed
In third-party secondary index;
3. in space expense, the secondary index itself based on coprocessor there are on HBase, HBase bottom be with
The storage of HFile format, HFile are compressed and are stored on HDFS, facilitate to save hard disk relative to third-party secondary index.
Detailed description of the invention
Fig. 1 is the flow chart that insertion module is inserted into the present invention.
Fig. 2 is the flow chart that interrogation model is inquired in the present invention.
Specific embodiment
With reference to the accompanying drawing, technical solution of the present invention is specifically described.
The present invention proposes a kind of HBase secondary index creation method based on coprocessor, according to pre- subregion and scattered at random
Index data and master data are carried out logical separation by column strategy;According to the different column families in same table, by index data and main number
According to progress physical separation.
In the present embodiment, secondary index is logically divided on the same Region according to pre- subregion and random hash
Area and main data area, secondary index area are used to store index data, and main data area is used to store master data.
In the present embodiment, same table is divided into two column families, a column family is used to store index data, another column
Race is used to master data.
It further, further include a kind of HBase secondary index creation system based on coprocessor, comprising:
It is inserted into module, for secondary index being constructed according to index configurations file, index data being inserted into data insertion
To secondary index area, newly-generated master data is inserted into main data area;
Enquiry module, for constructing querying condition, parallel query Region according to index configurations file in data query
Secondary index area obtain index line unit after, the asynchronous acquisition master data on Region.
Further, in the present embodiment, as shown in Figure 1, insertion module is indexed insertion in accordance with the following steps:
Step 11: when building table, calculating the number of Region, pre- subregion is carried out to each Region;
Step 12: after pre- subregion, obtaining the section startKey and endKey of each Region;
Step 13: rewriteeing prePut () Hook Function of coprocessor, master data is inserted by function acquisition
rowkey;
Step 14: a value in the section startKey and endKey of Region is randomly generated, and is denoted as hashKey;
By hashKey splicing before being inserted into the rowkey of master data, and remember that the line unit being newly inserted into is hashRowkey;
Step 15: the search index configuration file in prePut () Hook Function, after obtaining the field indexed, parsing
Put object judges whether there is index field insertion;If so, the startKey in the section Region where hashKey is then obtained, and
The value of index field is spliced before being inserted into the line unit of master data, as index line unit;Otherwise directly put object is inserted into and is led
Data field;
Step 16: obtaining starting line unit, index train value and be inserted into the length value of line unit, splice according to following format: rising
Begin key, and _ index train value _ is inserted into line unit, and is inserted into a column of index column family;
Step 17: index data being inserted into index column family, master data is inserted into data column family;The title of index column family is solid
Fixed, do not allow to modify and do not allow to weigh with index column Praenomen name namely unchangeable while also not allowing other column
Praenomen is known as index.
Further, in the present embodiment, in step s 11, further include following steps:
Step S111: determine that the pre- subregion number of cluster, the pre- subregion number calculation formula of individual node are as follows:
Wherein, M indicates the memory size of RegionServer;F indicates that RegionServer gives the ratio of memstore,
0.4 is defaulted as in HBase;S indicates the size of memstore, and the default value in unit M, HBase is 128;A is column family in table
Number;In the present embodiment, table includes at least 2 column families: one be storage secondary index column family index, it is another
A column family data for storage master data;
Step S112: determine that the node number of cluster, the calculation formula of the total pre- subregion number of cluster are as follows:
R=P*N
Wherein, R indicates the total number of the pre- subregion of cluster, and P indicates the number of the pre- subregion of each node, and N indicates to save in cluster
The number of point.
Further, in the present embodiment, in step s 13, further include following steps:
Step S131: client issues put request;
Step S132: the request is assigned to corresponding suitable RegionServer and Region;
Step S133: coprocessor intercepts the request, then on each RegionObserver online on the table
PrePut () Hook Function is called, put request is intercepted, parses put object, obtain rowkey value;
Step S134: if do not intercepted by prePut () Hook Function, put request continues to be sent to Region, then
It is handled.
Further, in the present embodiment, as shown in Fig. 2, enquiry module includes the following steps is inquired:
Step 21: when querying condition is arranged in client, querying condition being parsed by enquiring component, reads index configurations text
Part judges that the field whether there is in index configurations file;
Step 22: if it does, indicating that the field establishes index, then constructing querying condition by search index device
It originates line unit and terminates line unit;The startKey that the starting line unit of querying condition is each Region splices querying condition, looks into
The startKey that line unit is each Region that terminates of inquiry condition splices a value just greater than querying condition, and goes to step
Rapid S24;
Step 23: if it does not exist, then carrying out full table scan, and going to step S27;
Step 24: after constructing new querying condition, the secondary index area in multi-thread concurrent to each Region is carried out
Scan inquiry;
Step 25: inquiring qualified index line unit in secondary index area, this is then parsed by the train value indexed
The line unit of the master data of record;
Step 26: asynchronous to obtain master data in batches on the Region after the index line unit for finding a Region
Record;
Step 27: result being pooled at enquiring component, then is pooled to client return.
The above are preferred embodiments of the present invention, all any changes made according to the technical solution of the present invention, and generated function is made
When with range without departing from technical solution of the present invention, all belong to the scope of protection of the present invention.
Claims (8)
1. a kind of HBase secondary index creation method based on coprocessor, which is characterized in that according to pre- subregion and random hash
Index data and master data are carried out logical separation by strategy;According to the different column families in same table, by index data and master data
Carry out physical separation.
2. a kind of HBase secondary index creation method based on coprocessor according to claim 1, which is characterized in that
According to pre- subregion and random hash, on the same Region, it is logically divided into secondary index area and main data area, secondary index
Area is used to store index data, and main data area is used to store master data.
3. a kind of HBase secondary index creation method based on coprocessor according to claim 1, which is characterized in that
Same table is divided into two column families, a column family is used to store index data, another column family is used to master data.
4. a kind of HBase secondary index based on coprocessor creates system characterized by comprising
It is inserted into module, for secondary index being constructed according to index configurations file, index data being inserted into two in data insertion
Grade index area, is inserted into main data area for newly-generated master data;
Enquiry module constructs querying condition according to index configurations file in data query, and the two of parallel query Region
After grade index area obtains index line unit, the asynchronous acquisition master data on Region.
5. a kind of HBase secondary index based on coprocessor according to claim 4 creates system, which is characterized in that
The insertion module is indexed insertion in accordance with the following steps:
Step 11: when building table, calculating the number of Region, pre- subregion is carried out to each Region;
Step 12: after pre- subregion, obtaining the section startKey and endKey of each Region;
Step 13: rewriteeing prePut () Hook Function of coprocessor, the rowkey for being inserted into master data is obtained by the function;
Step 14: a value in the section startKey and endKey of Region is randomly generated, and is denoted as hashKey;It should
HashKey splices before being inserted into the rowkey of master data, and remembers that the line unit being newly inserted into is hashRowkey;
Step 15: the search index configuration file in prePut () Hook Function parses put after obtaining the field indexed
Object judges whether there is index field insertion;If so, then obtaining the startKey in the section Region where hashKey, and will
The value of index field is spliced before being inserted into the line unit of master data, as index line unit;Otherwise put object is directly inserted into main number
According to area;
Step 16: obtaining starting line unit, index train value and be inserted into the length value of line unit, splice according to following format: initial row
Key _ index train value _ is inserted into line unit, and is inserted into a column of index column family;
Step 17: index data being inserted into index column family, master data is inserted into data column family;It is fixed for indexing the title of column family
, do not allow to modify and do not allow name of weighing with index column Praenomen.
6. a kind of HBase secondary index based on coprocessor according to claim 5 creates system, which is characterized in that
Further include following steps in the step S11:
Step S111: determine that the pre- subregion number of cluster, the pre- subregion number calculation formula of individual node are as follows:
Wherein, M indicates the memory size of RegionServer;F indicates that RegionServer gives the ratio of memstore;S is indicated
The size of memstore, unit M;A is the number of column family in table;
Step S112: determine that the node number of cluster, the calculation formula of the total pre- subregion number of cluster are as follows:
R=P*N
Wherein, R indicates the total number of the pre- subregion of cluster, and P indicates the number of the pre- subregion of each node, and N indicates cluster interior joint
Number.
7. a kind of HBase secondary index based on coprocessor according to claim 5 creates system, which is characterized in that
Further include following steps in the step S13:
Step S131: client issues put request;
Step S132: the request is assigned to corresponding RegionServer and Region;
Step S133: coprocessor intercepts the request, then calls on each RegionObserver online on the table
PrePut () Hook Function intercepts put request, parses put object, obtains rowkey value;
Step S134: if do not intercepted by prePut () Hook Function, put request continues to be sent to Region, then carries out
Processing.
8. a kind of HBase secondary index based on coprocessor according to claim 4 creates system, which is characterized in that
The enquiry module, which includes the following steps, to be inquired:
Step 21: when querying condition is arranged in client, querying condition is parsed by enquiring component, reads index configurations file,
Judge that the field whether there is in index configurations file;
Step 22: if it does, indicating that the field establishes index, then constructing the starting of querying condition by search index device
Line unit and end line unit;The startKey that the starting line unit of querying condition is each Region splices querying condition, inquires item
The startKey that line unit is each Region that terminates of part splices a value greater than querying condition, and goes to step S24;
Step 23: if it does not exist, then carrying out full table scan, and going to step S27;
Step 24: after constructing new querying condition, the secondary index area in multi-thread concurrent to each Region carries out primary
Scan inquiry;
Step 25: inquiring qualified index line unit in secondary index area, the record is then parsed by the train value indexed
Master data line unit;
Step 26: after the index line unit for finding a Region, the asynchronous record for obtaining master data in batches on the Region;
Step 27: result being pooled at enquiring component, then is pooled to client return.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810945470.3A CN109165222A (en) | 2018-08-20 | 2018-08-20 | A kind of HBase secondary index creation method and system based on coprocessor |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810945470.3A CN109165222A (en) | 2018-08-20 | 2018-08-20 | A kind of HBase secondary index creation method and system based on coprocessor |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109165222A true CN109165222A (en) | 2019-01-08 |
Family
ID=64895931
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810945470.3A Pending CN109165222A (en) | 2018-08-20 | 2018-08-20 | A kind of HBase secondary index creation method and system based on coprocessor |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109165222A (en) |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109977113A (en) * | 2019-01-25 | 2019-07-05 | 北京工业大学 | A kind of HBase Index Design method based on Bloom filter for medical imaging data |
CN111046074A (en) * | 2019-12-13 | 2020-04-21 | 北京百度网讯科技有限公司 | Streaming data processing method, device, equipment and medium |
CN111143363A (en) * | 2019-12-23 | 2020-05-12 | 武汉光谷信息技术股份有限公司 | 3D Tiles data access method and device based on HBase |
CN111427887A (en) * | 2020-03-17 | 2020-07-17 | 中国邮政储蓄银行股份有限公司 | Method, device and system for rapidly scanning HBase partition table |
CN111680043A (en) * | 2020-06-05 | 2020-09-18 | 南京莱斯信息技术股份有限公司 | Method for rapidly searching mass data |
CN111737267A (en) * | 2020-08-03 | 2020-10-02 | 深圳市赢时胜信息技术股份有限公司 | HBase-based index system and query acceleration method |
CN112084188A (en) * | 2020-08-25 | 2020-12-15 | 北京明略昭辉科技有限公司 | HBase memory index construction method, system and storage medium |
CN112765131A (en) * | 2021-01-22 | 2021-05-07 | 重庆邮电大学 | Heterogeneous medical health data storage and retrieval method and system |
CN113032479A (en) * | 2019-12-24 | 2021-06-25 | 上海昂创信息技术有限公司 | HBase non-primary key indexing method and HBase system |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104376053A (en) * | 2014-11-04 | 2015-02-25 | 南京信息工程大学 | Storage and retrieval method based on massive meteorological data |
CN105787118A (en) * | 2016-03-25 | 2016-07-20 | 武汉工程大学 | Design method and query method for HBase secondary index |
-
2018
- 2018-08-20 CN CN201810945470.3A patent/CN109165222A/en active Pending
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104376053A (en) * | 2014-11-04 | 2015-02-25 | 南京信息工程大学 | Storage and retrieval method based on massive meteorological data |
CN105787118A (en) * | 2016-03-25 | 2016-07-20 | 武汉工程大学 | Design method and query method for HBase secondary index |
Non-Patent Citations (1)
Title |
---|
张威: ""环境空气质量监测大数据非侵入式二级索引的研究"", 《中国优秀硕士学位论文全文数据库 (信息科技辑)》 * |
Cited By (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109977113A (en) * | 2019-01-25 | 2019-07-05 | 北京工业大学 | A kind of HBase Index Design method based on Bloom filter for medical imaging data |
CN111046074A (en) * | 2019-12-13 | 2020-04-21 | 北京百度网讯科技有限公司 | Streaming data processing method, device, equipment and medium |
CN111046074B (en) * | 2019-12-13 | 2023-09-01 | 北京百度网讯科技有限公司 | Streaming data processing method, device, equipment and medium |
CN111143363A (en) * | 2019-12-23 | 2020-05-12 | 武汉光谷信息技术股份有限公司 | 3D Tiles data access method and device based on HBase |
CN113032479A (en) * | 2019-12-24 | 2021-06-25 | 上海昂创信息技术有限公司 | HBase non-primary key indexing method and HBase system |
CN111427887A (en) * | 2020-03-17 | 2020-07-17 | 中国邮政储蓄银行股份有限公司 | Method, device and system for rapidly scanning HBase partition table |
CN111680043A (en) * | 2020-06-05 | 2020-09-18 | 南京莱斯信息技术股份有限公司 | Method for rapidly searching mass data |
CN111680043B (en) * | 2020-06-05 | 2023-11-28 | 南京莱斯信息技术股份有限公司 | Method for quickly retrieving mass data |
CN111737267B (en) * | 2020-08-03 | 2021-01-26 | 深圳市赢时胜信息技术股份有限公司 | HBase-based index system and query acceleration method |
CN111737267A (en) * | 2020-08-03 | 2020-10-02 | 深圳市赢时胜信息技术股份有限公司 | HBase-based index system and query acceleration method |
CN112084188A (en) * | 2020-08-25 | 2020-12-15 | 北京明略昭辉科技有限公司 | HBase memory index construction method, system and storage medium |
CN112765131A (en) * | 2021-01-22 | 2021-05-07 | 重庆邮电大学 | Heterogeneous medical health data storage and retrieval method and system |
CN112765131B (en) * | 2021-01-22 | 2023-03-24 | 重庆邮电大学 | Heterogeneous medical health data storage and retrieval method and system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109165222A (en) | A kind of HBase secondary index creation method and system based on coprocessor | |
CN104536959B (en) | A kind of optimization method of Hadoop accessing small high-volume files | |
CN106663056B (en) | Metadata index search in a file system | |
US7523130B1 (en) | Storing and retrieving objects on a computer network in a distributed database | |
CN104794123B (en) | A kind of method and device building NoSQL database indexes for semi-structured data | |
CN102122285B (en) | Data cache system and data inquiry method | |
JP5006472B2 (en) | Table search device, table search method, and table search system | |
US20100287166A1 (en) | Method and system for search engine indexing and searching using the index | |
CN107368527B (en) | Multi-attribute index method based on data stream | |
CN104850572A (en) | HBase non-primary key index building and inquiring method and system | |
CN108228799B (en) | Object index information storage method and device | |
CN104572983A (en) | Construction method based on hash table of memory, text searching method and corresponding device | |
US11281645B2 (en) | Data management system, data management method, and computer program product | |
Pei et al. | An efficient query scheme for hybrid storage blockchains based on merkle semantic trie | |
WO2021006977A1 (en) | Delta graph traversing system | |
CN113254535A (en) | Method and device for synchronizing data from mongodb to mysql and computer readable storage medium | |
US9607044B2 (en) | Systems and methods for searching multiple related tables | |
CN108241709B (en) | Data integration method, device and system | |
CN104636368A (en) | Data retrieval method and device and server | |
CN111582967A (en) | Content search method, device, equipment and storage medium | |
CN100357943C (en) | A method for inspecting garbage files in cluster file system | |
CN116756253A (en) | Data storage and query methods, devices, equipment and media of relational database | |
CN106776810A (en) | The data handling system and method for a kind of big data | |
US10185742B2 (en) | Flexible text searching for data objects of object notation | |
CN102799996A (en) | Network advertisement strategy matching method and system |
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 | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190108 |