This operator represents the fetching of columns from a table.
Operator name: FETCH
Represents: The fetching of columns from a table using a specific row identifier (RID).
- Expand index keys to include the fetched columns so that the data pages do not have to be accessed.
- Find the index related to the fetch, and double-click on its node to display its statistics window. Ensure that the degree of clustering is high for the index.
- Increase the buffer size if the input/output (I/O) incurred by the fetch is greater than the number of pages in the table.
- If statistics are not current, update them.
The quantile and frequent value statistics provide information on the selectivity of predicates, which determines when index scans are chosen over table scans. To update these statistics, use the RUNSTATS command on a table with the WITH DISTRIBUTION clause.