Richard Huxton <dev@xxxxxxxxxxxx> writes: > soni de wrote: >> NOTICE: QUERY PLAN: >> Sort (cost=17.13..17.13 rows=1 width=16) (actual time=619140.18..619140.29rows >> =288 loops=1) >> -> Index Scan using lan_pkey on lan (cost=0.00..17.12 rows=1 width=16) >> (actual time=7564.44..619121.61 rows=288 loops=1) >> >> Total runtime: 619140.76 msec > OK - there is clearly something wrong here when you take 10 minutes to > fetch 288 rows from an index. > I'm guessing you have a *lot* of dead rows in there. Yeah. The other small problem here is that EXPLAIN output hasn't looked like that since PG 7.2 (unless Soni has just omitted the index-condition lines). I'd recommend updating to something modern. regards, tom lane
