Re: [HACKERS] LIKE fixed(?) for non-ASCII collation orders

Yes, we always still match the LIKE. The additional comparisons just
make the number of rows smaller.

> Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> writes:
>
> > I have just committed what I hope is the final solution for the problem
> > of LIKE index optimization in non-ASCII locales. indxpath.c now
> > generates both a lower and upper indexqual in all locales. For example,
> > x LIKE 'foo%t'
> > will create indexqual conditions
> > x >= 'foo' AND x < 'fop'
> > The "<" condition is omitted only if the code is unable to produce a
> > string greater than the pattern's constant prefix.
>
> the .. >= .. < .. condition will result in addtional matches, like 'fo ot',
> so you still have to check with LIKE. I'm using such an expression
> (without the additional LIKE) in an application, and it seems to match
> at least everything that LIKE would match too (my users would have
> complained about missing matches, but i never did a formal test or
> evaluation).
>
> ************
>

--
Bruce Momjian | http://www.op.net/~candle
maillist(at)candle(dot)pha(dot)pa(dot)us | (610) 853-3000
+ If your life is a hard drive, | 830 Blythe Avenue
+ Christ can be your backup. | Drexel Hill, Pennsylvania 19026