On Fri, 25 Oct 2002, [iso-8859-15] Herv� Piedvache wrote:
> Thanks for you help ... to "explain" us why Postgresql do not use all the
> time the index scan for this kind of request for 3 tables with the same
> primary key id.
Depending on the number of rows returned, index scan may not be faster
than a sequential scan. As you up the constant, the estimated number of
rows drops which make the index scan a better and better plan. Are
the estimated number of rows for the scan of a (4379, 3987, and 29)
realistic for mynumber>(23000, 23500, 24000)