-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

> In practice, I don't think that LIKE-style patterns (% and _ wildcards)
> will pose a serious compatibility problem if we just decree that the
> -n and -t switches now take patterns rather than plain names.  I agree
> that regex-style patterns would open some gotchas, but what's wrong with
> standardizing on LIKE patterns?

Sounds good, but the more I think about it, why don't we just use regexes via
the ~ operator? After all, if we want to exclude schemas starting with an
underscore from pg_dump, then -N '^_.*' is no worse than -N '\\_%' and has
the added advantage of being more like regexes people are used to. I guess
my earlier 'which is which' argument isn't too much to worry about either -
chances are very slim that an existing script is using a -t argument that
contains regular expressions. Plus, while the underscore is common in
namespace names, a period is not.

- --
Greg Sabino Mullane [EMAIL PROTECTED]
PGP Key: 0x14964AC8 200601172005
http://biglumber.com/x/web?pk=2529DF6AB8F79407E94445B4BC9B906714964AC8
-----BEGIN PGP SIGNATURE-----

iD8DBQFDzZSpvJuQZxSWSsgRAiEQAKD5YXJjne5ZjbSUyHLiVKrEBtLPxQCfbsN8
JlQH5S+UVTogKpyRQJoU6jk=
=Sfcu
-----END PGP SIGNATURE-----



---------------------------(end of broadcast)---------------------------
TIP 9: In versions below 8.0, the planner will ignore your desire to
       choose an index scan if your joining column's datatypes do not
       match

Reply via email to