DATABASE QUERY IN A SHARE-NOTHING DATABASE ARCHITECTURE

• US 20140095470A1
• Filed: 08/13/2013
• Published: 04/03/2014
• Est. Priority Date: 09/29/2012
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Chen, Xing, Jin, Yi, Zhang, Wan Chuan, Hao, Qing Yun
• Granted Patent: US 9,146,959 B2
• Time in Patent Office: Days
• US Class Current: 707/714
• CPC Class Codes: G06F 16/24524 Access plan code generation and invalidation; Reuse of access plans
• CPC Class Codes: G06F 16/24534 Query rewriting; Transformation
• CPC Class Codes: G06F 16/24542 Plan optimisation
• CPC Class Codes: G06F 16/24544 Join order optimisation

Abstract

Data query in a share-nothing database includes obtaining a query request and generating an optimized access plan with respect to the query request. The query request relates to external data stored in an external data source and contains a definition for expected distribution of the external data. The data query also includes obtaining data distribution information related to the expected distribution based on the optimized access plan, transmitting the data distribution information to the external data source so that the external data source splits and returns the external data in accordance with the data distribution information, and executing query-related processing of the split external data in accordance with the optimized access plan.

Claims

1. A method for data query in a share-nothing database, comprising:

obtaining, by a computer processor, a query request and generating an optimized access plan with respect to the query request, wherein the query request relates to external data stored in an external data source, and the query request further contains a definition for expected distribution of the external data;

obtaining data distribution information related to the expected distribution of the external data based on the optimized access plan;

transmitting the data distribution information to the external data source so that the external data source splits the external data in accordance with the data distribution information and returns the split external data in parallel; and

executing query-related processing on the split external data in accordance with the optimized access plan.

2. The method according to claim 1, wherein the query request further relates to internal data stored in the share-nothing database, and comprises an association condition of the external data and internal data, wherein the expected distribution of the external data is associated with the association condition.

3. The method according to claim 1, wherein the generating an optimized access plan with respect to the query request comprises:

recognizing the definition for the expected distribution of the external data, and including it in a preliminary access plan; and

optimizing the preliminary access plan considering the expected distribution to obtain the optimized access plan.

4. The method according to claim 1, wherein the obtaining the data distribution information comprises:

obtaining information of the optimized access plan by calling an extended interface function in runtime, and obtaining the data distribution information based on the obtained information of the optimized access plan.

5. The method according to claim 4, wherein the data distribution information comprises a direct splitting rule and returning rule, so that the external data source splits the external data in accordance with the splitting rule and returns the split external data in accordance with the returning rule.

6. The method according to claim 4, wherein the data distribution information relates to index information pointing to other data, so that the external data source first obtains contents of the other data in accordance with the index information, accordingly determines a target distribution of the external data, and then splits and returns the external data based on the target distribution of the external data.

7. The method according to claim 4, wherein the data distribution information relates to index information pointing to other data, and the transmitting the data distribution information to the external data source comprises firstly processing the data distribution information so that it contains a direct splitting rule and returning rule, and then transmitting the processed data distribution information to the external data source.

8. The method according to claim 2, wherein the executing query-related processing on the external data comprises:

executing co-located JOIN on the internal data and the split external data, the co-located JOIN referring to JOIN within a same node in the share-nothing database.

9. An apparatus for data query in a share-nothing database, comprising:

an access plan generation unit configured to obtain a query request and generate an optimized access plan with respect to the query request, wherein the query request relates to external data stored in an external data source, and the query request further contains a definition for expected distribution of the external data;

a data distribution information obtaining unit configured to obtain data distribution information related to the expected distribution of the external data based on the optimized access plan;

a data distribution information transmission unit configured to transmit the data distribution information to the external data source so that the external data source splits the external data in accordance with the data distribution information and returns the split external data in parallel; and

a query processing execution unit configured to execute query-related processing on the split external data in accordance with the optimized access plan.

10. The apparatus according to claim 9, wherein the query request further relates to internal data stored in the share-nothing database, and comprises an association condition of the external data and internal data, wherein the expected distribution of the external data is associated with the association condition.

11. The apparatus according to claim 9, wherein the access plan generation unit is configured to:

recognize the definition for the expected distribution of the external data, and include it in a preliminary access plan; and

optimize the preliminary access plan considering the expected distribution to obtain the optimized access plan.

12. The apparatus according to claim 9, wherein the data distribution information obtaining unit is configured to:

obtain information of the optimized access plan by calling an extended interface function in runtime, and obtain the data distribution information based on the obtained information of the optimized access plan.

13. The apparatus according to claim 12, wherein the data distribution information comprises a direct splitting rule and returning rule, so that the external data source splits the external data in accordance with the splitting rule and returns the split external data in accordance with the returning rule.

14. The apparatus according to claim 12, wherein the data distribution information relates to index information pointing to other data, so that the external data source first obtains contents of the other data in accordance with the index information, accordingly determines a target distribution of the external data, and then splits and returns the external data based on the target distribution of the external data.

15. The apparatus according to claim 12, wherein the data distribution information obtained by the data distribution information obtaining unit relates to index information pointing to other data, and the data distribution information transmission unit is configured to firstly process the data distribution information so that it contains a direct splitting rule and returning rule, and then transmit the processed data distribution information to the external data source.

16. The apparatus according to claim 10, wherein the query processing execution unit is configured to execute co-located JOIN on the internal data and the split external data, the co-located JOIN referring to JOIN within a same node in the share-nothing database.