Tatsuo Ishii <t-ishii@sra.co.jp> writes:
> Here is a new contrib function called "pgstatindex", similar to
> pgstattuple but different in that it returns the percentage of the
> dead tuples of an index. I am posting this for review purpose.
Um ... what's the point? Isn't this always the same as the percentage
for the underlying table?
regards, tom lane