Re: [BUGS] BUG #3829: Wrong index reporting from pgAdmin III (v1.8.0 rev 6766-6767)

Pavel Stehule wrote:
> Hello
> 
> it's pgAdmin bug. please report it there

Actually it doesn't appear to be...

Using the OPs example code on 8.3 beta 4:

>> CREATE INDEX test_desc_idx1
>>   ON test_desc
>>   USING btree
>>   (f1, f2 desc)
>>
>> CREATE INDEX test_desc_idx2
>>   ON test_desc
>>   USING btree
>>   (f1 desc, f2);

postgres=# select pg_get_indexdef('test_desc_idx1'::regclass, 0, true);                         pg_get_indexdef
--------------------------------------------------------------------CREATE INDEX test_desc_idx1 ON test_desc USING
btree(f1, f2 DESC)
 
(1 row)

postgres=# select pg_get_indexdef('test_desc_idx1'::regclass, 1, true);pg_get_indexdef
-----------------f1 DESC
(1 row)

postgres=# select pg_get_indexdef('test_desc_idx1'::regclass, 2, true);pg_get_indexdef
-----------------f2 DESC
(1 row)

postgres=# select pg_get_indexdef('test_desc_idx2'::regclass, 0, true);                         pg_get_indexdef
--------------------------------------------------------------------CREATE INDEX test_desc_idx2 ON test_desc USING
btree(f1 DESC, f2)
 
(1 row)

postgres=# select pg_get_indexdef('test_desc_idx2'::regclass, 1, true);pg_get_indexdef
-----------------f1 DESC
(1 row)

postgres=# select pg_get_indexdef('test_desc_idx2'::regclass, 2, true);pg_get_indexdef
----------------- DESCf2
(1 row)

Looks like pg_get_indexdef is unwell :-(

/D