Re: [HACKERS] Index creation takes for ever

On Sun, 07 Sep 2003 12:23:28 -0400, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
wrote:
>Maybe so, but it would degrade the performance in the unique-index case
>if we do it as the TODO is worded.

The patch would only hurt with a unique index, if there are lots of
duplicate tuples at CREATE INDEX time.

>My own opinion is that the bitmap-index-lookup approach will be superior

So is mine, but I was not able to do this in 30 lines. Sorry ;-)

>to trying to keep the index entries in TID order.
^^^^
... which the patch does not. I see its main advantage in creating
better b-tree indices when you restore a large database with many
duplicate index entries.

It is intended to be only effective at CREATE INDEX or REINDEX time.
I don't believe it is activated when you insert a single new entry,
otherwise it wouldn't pass regression tests ...

Servus
Manfred