Re: [HACKERS] Oracle Compatibility (Translate function) - Mailing list pgsql-hackers

From Thomas Lockhart
Subject Re: [HACKERS] Oracle Compatibility (Translate function)
Date
Msg-id 385A7135.51ECBBA2@alumni.caltech.edu
Whole thread Raw
In response to Re: [HACKERS] Re: [BUGS] uniqueness not always correct  (Bruce Momjian <pgman@candle.pha.pa.us>)
List pgsql-hackers
> I have modified the translate function in order to improve its
> compatibility with Oracle.  It now supports the replacement of 
> multiple characters and it will also shorten the length of the string 
> when characters are replaced with nothing.
> [Note: The arguments are different from the original translate]
> Can this function replace the existing function in the distribution?

afaik yes. Does anyone have a problem with this (it allows
substitution of multiple characters)? I think the system tables will
need to be updated; I'll do this within the next week or so if noone
else has already taken this on.

btw, there is some chance that when we go to native support for
NATIONAL CHARACTER etc then TRANSLATE() will need to become SQL92
compliant (and basically a different function). But that is an issue
for later, and we may be able to solve it without having to give up on
the Oracle version.
                 - Thomas

-- 
Thomas Lockhart                lockhart@alumni.caltech.edu
South Pasadena, California


pgsql-hackers by date:

Previous
From: Bruce Momjian
Date:
Subject: initdb.sh fixed
Next
From: Tom Lane
Date:
Subject: Re: [HACKERS] Anyone for prettyprinted EXPLAIN VERBOSE?