In Oracle DATABASE there is something called Cost Based OPTIMIZER (CBO) which other than doing many things also does query optimization. That is it determines the best or optimum method to run a query. Now when large changes to the data or database take place, the system performance may get slow. DBMS_STATS is the package which COLLECTS statistics for Cost Based Optimizer. Collecting Statistics lets the CBO know about the best methods for query optimization. The statistics include the Table statistics, Column statistics, Index statistics and System statistics. The DBMS_STATS package INCLUDES the following procedures.

1. GATHER_INDEX_STATS, Index statistics
2. GATHER_TABLE_STATS, Table, column, and index statistics
3. GATHER_SCHEMA_STATS,Statistics for all OBJECTS in a schema
4. GATHER_DICTIONARY_STATS,Statistics for all dictionary objects
5. GATHER_DATABASE_STATS,Statistics for all objects in a database 

Of all of the above usually the most used is GATHER_SCHEMA_STATS to gather statistics of all objects within a particular Schema. The Gather Schema Statistics process should be run on a regular basis (weekly at a minimum) and anytime large changes to the data or database take place.

In Oracle Database there is something called Cost Based Optimizer (CBO) which other than doing many things also does query optimization. That is it determines the best or optimum method to run a query. Now when large changes to the data or database take place, the system performance may get slow. DBMS_STATS is the package which collects statistics for Cost Based Optimizer. Collecting Statistics lets the CBO know about the best methods for query optimization. The statistics include the Table statistics, Column statistics, Index statistics and System statistics. The DBMS_STATS package includes the following procedures.

Of all of the above usually the most used is GATHER_SCHEMA_STATS to gather statistics of all objects within a particular Schema. The Gather Schema Statistics process should be run on a regular basis (weekly at a minimum) and anytime large changes to the data or database take place.