Updating statistics after upgrading

The statistical data located in sysstatistics is upgraded during the data copy section of upgrade, which can cause rows to expand beyond the server’s page size. However, before these rows are inserted during the data copy, large statistical rows are split in two. The statistics for previously unpartitioned tables are used as statistics for the resulting partitioned tables after the upgrade, and an object’s ID is used as its partition ID in the sysstatistics..partitionid column.

Sybase strongly recommends that you run update statistics after you upgrade, particularly if you have not recently run update statistics on some tables. Because Adaptive Server 15.0 has several algorithms for sorting, grouping, unions, joining, and other operations, it is must have current statistics. Earlier versions of Adaptive Server had a single algorithm, and it was not necessary to run update statistics on fairly static data, such as data in reporting systems. However, Adaptive Server 15.0 may select a slower algorithm if it does not have current statistics, as the actual data volume far exceeds the projected volume based on the stale statistics.

WARNING! Making unneeded changes to statistics may adversely affect performance. When you update statistics, consider using a larger step count than the default, particularly for large tables and when you change the histogram tuning factor. Doing this ensures that statistics on indexes on columns that contain a large number of duplicate values contain a more accurate representation of the data skew. Running update statistics with incorrect parameters may change the number of steps in a histogram and impact performance. Apply changes first in a test environment. See the Performance and Tuning Guide: Monitoring and Analyzing more for information.