"Michael G. Martin" <[EMAIL PROTECTED]> writes:
> Here is what I would expect which usually happens:
> explain select * from symbol_data where symbol_name='IBM';
> Index Scan using symbol_data_pkey on symbol_data (cost=0.00..512.99 rows=128
>width=129)
> Here is one that fails:
> explain select * from symbol_data where symbol_name='ELTE';
> Seq Scan on symbol_data (cost=0.00..707415.32 rows=438015 width=129)
The planner thinks that there are 438K occurrences of 'ELTE' in your
table. If that's true, a sequential scan is not obviously the wrong
choice. How many are there, in reality? What do you get from
select * from pg_stats where tablename = 'symbol_data' and attname =
'symbol_name';
Also, to put the rubber to the road: if you force an indexscan by
doing "set enable_seqscan = off", does it get faster or slower?
(EXPLAIN ANALYZE would be useful here.)
regards, tom lane
---------------------------(end of broadcast)---------------------------
TIP 2: you can get off all lists at once with the unregister command
(send "unregister YourEmailAddressHere" to [EMAIL PROTECTED])
