Erich Stamberger <eberger@gewi.kfunigraz.ac.at> writes:
> Another interresting feature of Czech collation is:
> H < "CH" < I
Oh my, that *is* interesting (using the word in the spirit of the
ancient Chinese curse, "May you live in interesting times"...)
> So what happens with "WHERE name like 'Czec%`" ?
The wrong thing, without doubt.
Would it help any to strip off one character of the given pattern?
That is, if the pattern is LIKE 'foo%', forget about the last 'o'
and generate bounds like 'fo' <= x <= 'fp' ?
regards, tom lane