Estimating statistics for generating execution plans for database queries

1. A computer-implemented method for generating statistics for processing database queries, the method comprising:

receiving, by a query compiler, from a client device, a plurality of database queries by a database system, the database queries processing data stored in database tables of the database system;

identifying, by the query compiler, missing statistics while generating execution plans for database queries, the identifying comprising, for each of the plurality of database queries:

requesting a statistical information useful for generating an execution plan for a database query; and

determining that the requested statistical information is not available; and

storing information describing the requested statistical information as missing statistics, responsive to determining that the requested statistical information is not available;

ranking the identified missing statistics based on a number of times each missing statistics was identified as being not available during generation of execution plan;

determining a subset of the identified missing statistics for use in generating execution plans for subsequent database queries, wherein determining the subset comprises selecting identified missing statistics based on the ranking;

receiving, by the query compiler, the subsequent database queries;

generating, by the query compiler, an execution plan for the one of the subsequent database queries using at least one of the identified missing statistics from the subset;

executing, by an execution engine, the generated execution plan for the one of the subsequent queries to determine a result set; and

sending the determined result set to the client device.

2. The computer-implemented method of claim 1 , wherein the requested statistical information is associated with one or more columns of database tables and wherein storing information describing the requested statistical information comprises storing information identifying the one or more columns.

3. The computer-implemented method of claim 1 , wherein the identified missing statistics is a multi-column number of distinct values (NDV) of a plurality of columns used in a database query.

4. The computer-implemented method of claim 1 , wherein identifying the missing statistics for a database query comprises:

determining that the database query processes a plurality of columns; and

responsive to determining that the database query processes a plurality of columns, identifying the statistical information to be a multi-column NDV of the plurality of columns.

5. The computer-implemented method of claim 4 , wherein the database query specifies a comparison of each column from the plurality of columns with a corresponding constant value.

6. The computer-implemented method of claim 4 , wherein the database query specifies a grouping of results of the database query based on distinct values of the plurality of columns.

7. The computer-implemented method of claim 4 , the database query specifies an equijoin operation based on columns from the plurality of columns.

8. The computer-implemented method of claim 1 , wherein determining the subset of the identified missing statistics comprises selecting missing statistics that was identified as missing for queries processed by a particular application.

9. The computer-implemented method of claim 1 , wherein determining the subset of the identified missing statistics comprises selecting missing statistics that was identified as missing for queries taking more than a threshold amount of time for processing.

10. The computer-implemented method of claim 1 , wherein determining the subset of the identified missing statistics comprises selecting missing statistics that was identified as missing for queries executed in a given time interval.

11. The computer-implemented method of claim 1 , wherein determining the subset of the identified missing statistics comprises selecting missing statistics that was identified as missing for queries associated with one of: a particular user, an application, or a work-load-management related identifier.

12. A non-transitory computer readable storage medium storing instructions thereon, the instructions for execution by a computer processor to cause the computer processor to perform:

receiving, by a query compiler from a client device, a plurality of database queries by a database system, the database queries processing data stored in database tables of the database system;

identifying, by the query compiler, missing statistics while generating execution plans for database queries, the identifying comprising, for each of the plurality of database queries:

requesting a statistical information useful for generating an execution plan for a database query;

determining that the requested statistical information is not available; and

storing information describing the requested statistical information as missing statistics, responsive to determining that the requested statistical information is not available;

ranking the identified missing statistics based on a number of times each missing statistics was identified as being not available during generation of execution plan;

determining a subset of the identified missing statistics for use in generating execution plans for subsequent database queries, wherein determining the subset comprises selecting identified missing statistics based on the ranking;

receiving, by the query compiler, the subsequent database queries;

generating, by the query compiler, an execution plan for the one of the subsequent database queries using at least one of the identified missing statistics from the subset;

executing, by an execution engine, the generated execution plan for the one of the subsequent queries to determine a result set; and

sending the determined result set to the client device.

13. The non-transitory computer readable storage medium of claim 12 , wherein the requested statistical information is associated with one or more columns of database tables and wherein storing information describing the requested statistical information comprises storing information identifying the one or more columns.

14. The non-transitory computer readable storage medium of claim 12 , wherein identifying the missing statistics for a database query comprises:

determining that the database query processes a plurality of columns; and

responsive to determining that the database query processes a plurality of columns, identifying the statistical information to be a multi-column NDV of the plurality of columns.

15. The non-transitory computer readable storage medium of claim 14 , wherein the database query specifies a comparison of each column from the plurality of columns with a corresponding constant value.

16. The non-transitory computer readable storage medium of claim 12 , wherein instructions for determining the subset of the identified missing statistics comprise instructions for selecting missing statistics that was identified as missing for database queries processed by a particular application.

17. The non-transitory computer readable storage medium of claim 12 , wherein instructions for determining the subset of the identified missing statistics comprise instructions for selecting missing statistics that was identified as missing for database queries associated with one of: a particular user, an application, a time window, or a work-load-management related identifier.

18. A computer-implemented system for generating statistics for a database system, the system comprising:

a computer processor; and

a computer-readable storage medium storing instructions thereon, the instructions for execution by a computer processor to cause the computer processor to perform:

receiving, by a query compiler, from a client device, a plurality of database queries by a database system, the database queries processing data stored in database tables of the database system;

identifying, by the query compiler, missing statistics while generating execution plans for database queries, the identifying comprising, for each of the plurality of database queries:

requesting a statistical information useful for generating an execution plan for a database query;

determining that the requested statistical information is not available; and

storing information describing the requested statistical information as missing statistics, responsive to determining that the requested statistical information is not available;

ranking the identified missing statistics based on a number of times each missing statistics was identified as being not available during generation of execution plan;

determining a subset of the identified missing statistics for use in generating execution plans for subsequent database queries, wherein determining the subset comprises selecting identified missing statistics based on the ranking;

receiving, by the query compiler, the subsequent database queries;

generating, by the query compiler, an execution plan for the one of the subsequent database queries using at least one of the identified missing statistics from the subset;

executing, by an execution engine, the generated execution plan for the one of the subsequent queries to determine a result set; and

sending the determined result set to the client device.

19. The computer-implemented system of claim 18 , wherein instructions for identifying the missing statistics for a database query comprise instructions for:

determining that the database query processes a plurality of columns; and

responsive to determining that the database query processes a plurality of columns, identifying the statistical information to be a multi-column NDV of the plurality of columns.

20. The computer-implemented system of claim 19 , wherein the database query specifies a comparison of each column from the plurality of columns with a corresponding constant value.