| Full statistics |
- Collects all statistics for all of the data.
- Time consuming.
- Most accurate of the three methods of collecting statistics.
- Stored in interval histograms in the data dictionary.
|
- Best choice for columns or indexes with highly skewed data values.
- Recommended for small tables, where a small table is one with fewer than 1,000 rows per AMP.
- Recommended for selection columns having a moderate to low number of distinct values.
- Recommended for most NUSIs and other selection columns.
Collection time on NUSIs is very fast.
- Recommended for all column sets or index column sets for the following cases.
Where full statistics add value.
Where sampling does not provide satisfactory statistical estimates.
|
| User-sampled statistics |
- Collects all statistics for a sample of the data, not just cardinality estimates.
- Significantly faster collection time than full statistics.
- Stored in interval histograms in the data dictionary.
|
|