Tom Lane wrote:
> which traces down to the fact that patternsel() expects to see only TEXT
> patterns. Joe, was it you that put this stuff in? If so, shame on you
> for not using --enable-cassert when doing backend work.
>
Shame already acknowledged -- you must have written this before reading
my last post. I know better now. Anyway, patch was sent in to patches.
A new problem exists though. Again see my post to patches. Indexed
lookups on bytea are completely broken on cvs tip. On 7.2.1:
test=# explain select * from bombytea where parent_part = 'FM04-13100-1';
NOTICE: QUERY PLAN:
Index Scan using bombytea_idx1 on bombytea (cost=0.00..148.37 rows=37
width=34)
EXPLAIN
test=# select * from bombytea where parent_part = 'FM04-13100-1';
parent_part | child_part | child_part_qty
--------------+------------+----------------
FM04-13100-1 | NULL | 0
(1 row)
and on cvs tip:
parts=# explain select * from bombytea where parent_part = 'FM04-13100-1';
QUERY PLAN
----------------------------------------------------------------------------------
Index Scan using bombytea_idx1 on bombytea (cost=0.00..147.83 rows=37
width=34)
Index Cond: (parent_part = 'FM04-13100-1'::bytea)
(2 rows)
parts=# select * from bombytea where parent_part = 'FM04-13100-1';
ERROR: Index bombytea_idx1 is not a btree
I'm trying to figure out why right now.
Joe