Improved distinct count estimates for skewed column data
Details
| Detail name | Value |
|---|---|
| Changelog Number | 10424 |
| Type | Improvement |
| Status | Resolved |
| Fix Versions | Exasol 7.0.1 |
| Resolution Date | 2020-09-23 |
Background
Column statistics (distinct count estimates, minimum & maximum values) are used by the join order optimizer to minimize pipeline costs. Distinct count estimates are especially important as the optimizer uses them to determine join and filter output cardinalities.
In case of heavily skewed column data, the distinct count estimates could be off from the real distinct count by several orders of magnitude.
Improvement
Distinct count estimates have been improved by using a new algorithm. The estimate has a much lower error and is independent of the order of values in a column.
