On Sun, 17 Jun 2001, Tom Lane wrote:
> * On large tables, ANALYZE uses a random sample of rows rather than
> examining every row, so that it should take a reasonably short time
> even on very large tables. Possible downside: inaccurate stats.
> We need to find out if the sample size is large enough.
How about letting the user specify how much of the table should be
examined? So if you know the data is homogenous you can just specify a
smaller sample than normal and if you have time you can have the whole
table scanned.
- Einar Karttunen