Postgresql - 慢“select * from table limit 1”查询

Question

I'm using PostgreSQL 9.3. I have database with one table, and this table have over 600 milions rows. When I connect to this database first query is really slow:

explain analyze select * from request_log limit 1;
Limit  (cost=0.00..0.02 rows=1 width=61) (actual time=481439.127..481439.129 rows=1 loops=1)
   ->  Seq Scan on request_log  (cost=0.00..13996870.79 rows=651159679 width=61) (actual time=481439.123..481439.123 rows=1 loops=1)
Total runtime: 481440.488 ms

I don't understand that - why seq scan just don't stop after reading first row?

Answer 1

I think that you mis-reading the output from explain analyse.

The line "Limit (cost=0.00..0.02 rows=1 width=61) (actual time=481439.127..481439.129 rows=1 loops=1)" shows that it has limited the sequential scan.

If you try it without the limit

EXPLAIN ANALYSE SELECT * FROM request_log;

you will (probably) see that it takes a lot longer.

The problem is something to do with data retrieval in your database. Check your postgresql.conf file under RESOURCE USUAGE and QUERY TUNING - look for anything unusual. Check you postgresql log for anything unusual.