Message-Id: <199912060935.UAA28036@mugca.cc.monash.edu.au>
Subject: Re: [INTERFACES] Problems with postgres V6.5.3 large objects
To: tgl@sss.pgh.pa.us (Tom Lane)
Date: Mon, 6 Dec 1999 20:35:00 +1100 (EST)
Cc: pgsql-interfaces@postgreSQL.org
In-Reply-To: <7702.944436768@sss.pgh.pa.us> from "Tom Lane" at Dec 5,
	99 06:32:48 pm
Reply-To: dougt@mugc.cc.monash.edu.au
From: Douglas Thomson <dougt@mugc.cc.monash.edu.au>
Content-Type: text

> Most of the developers have not wanted to put much effort into large
> objects, since where we really want to go is to eliminate tuple size
> restrictions; once that happens large objects will be much less
> necessary.

I am curious about the planned interface once tuple size restrictions
are eliminated. I am using large objects, and want to make my
application port as painlessly as possible.

If *all* that happens is that tuple size and SQL command length
limits are eliminated, then presumably I will just use a TEXT
attribute and do a normal INSERT to store my large object? But will
that mean that I have to go right through my large objects and escape
all the nasty binary characters (such as single quotes) that are
illegal in an SQL string constant?

If there is some other list where this discussion is accessible then
please direct me!

Doug.

P.S. Did anyone ever comment about whether multi-table joins took
     advantage of individual table indexes on the joining attributes,
     or whether the join had to read in all the table data and sort it
     anyway? There was some trouble with duplicated postings at the
     time, and I may have missed something while I was purging the
     messages I had already read...