Tom Lane wrote:
>Richard Neill <rn214@hermes.cam.ac.uk> writes:
>
>
>>This ordering is perverse!
>>
>>
>
>No kidding.
>
>
>
>>No matter what the priority is of the
>>different characters, I cannot understand how the above can arise.
>>
>>
>
>You are assuming that it's a byte-at-a-time process. It's not. I
>believe the first pass considers only letters and digits.
>
>You can easily prove to yourself that it's not just Postgres. Here's
>an example on my Linux laptop:
>
>[tgl@g3 tgl]$ cat zzz
>Cymbal #1
>Cymbal - 18 inch
>Cymbal #2
>[tgl@g3 tgl]$ LC_ALL=C sort zzz
>Cymbal #1
>Cymbal #2
>Cymbal - 18 inch
>[tgl@g3 tgl]$ LC_ALL=en_GB sort zzz
>Cymbal #1
>Cymbal - 18 inch
>Cymbal #2
>[tgl@g3 tgl]$
>
>
I verified this, and it's not GB specific as one might suggest... Same
with en_US, de_DE, fr_FR, af_ZA. Does this behaviour really make sense
to anybody?
Regards,
Andreas
