Double ocurring Subplan - Mailing list pgsql-hackers

From Gurjeet Singh
Subject Double ocurring Subplan
Date
Msg-id BANLkTinqsWv5g2WOaokPoY6SXxUgF=Whwg@mail.gmail.com
Whole thread Raw
Responses Re: Double ocurring Subplan
List pgsql-hackers
I am listing the query, it's explain output and explain analyze output at the end.

The EXPLAIN output shows the 'Subplan 2' node only once, whereas EXPLAIN ANALYZE output shows that the 'Subplan 2' is being executed twice . Is that true? Or is it just the plan printer getting confused? Is the confusion because of the 2 conditions in the WHERE clause of the correlated subquery?

PG Version:
 PostgreSQL 9.0.3 on x86_64-unknown-linux-gnu, compiled by GCC gcc (Ubuntu/Linaro 4.4.4-14ubuntu5) 4.4.5, 64-bit

The query:

select d.m1 - h.m1
from tz_test as d
join tz_test_history as h
on d.server_id = h.server_id
where d.server_id = 5
and h.recorded_time = (select max(recorded_time)
                        from tz_test_history as h2
                        where h2.server_id = d.server_id
                        and h2.recorded_time < d.recorded_time);

The explain output:
                                                                   QUERY PLAN                                                                  
------------------------------------------------------------------------------------------------------------------------------------------------
 Nested Loop  (cost=2.26..15.54 rows=1 width=8)
   ->  Seq Scan on tz_test d  (cost=0.00..1.62 rows=1 width=16)
         Filter: (server_id = 5)
   ->  Index Scan using tz_test_hist_rec_time_idx on tz_test_history h  (cost=2.26..11.64 rows=1 width=16)
         Index Cond: (h.recorded_time = (SubPlan 2))
         Filter: (h.server_id = 5)
         SubPlan 2
           ->  Result  (cost=2.25..2.26 rows=1 width=0)
                 InitPlan 1 (returns $2)
                   ->  Limit  (cost=0.00..2.25 rows=1 width=8)
                         ->  Index Scan Backward using tz_test_hist_rec_time_idx on tz_test_history h2  (cost=0.00..10800.38 rows=4800 width=8)
                               Index Cond: ((recorded_time IS NOT NULL) AND (recorded_time < $1))
                               Filter: (server_id = $0)


explain analyze output:

                                                                                        QUERY PLAN
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
 Nested Loop  (cost=2.26..15.54 rows=1 width=8) (actual time=0.236..0.289 rows=1 loops=1)
   ->  Seq Scan on tz_test d  (cost=0.00..1.62 rows=1 width=16) (actual time=0.031..0.072 rows=1 loops=1)
         Filter: (server_id = 5)
   ->  Index Scan using tz_test_hist_rec_time_idx on tz_test_history h  (cost=2.26..11.64 rows=1 width=16) (actual time=0.103..0.112 rows=1 loops=1)
         Index Cond: (h.recorded_time = (SubPlan 2))
         Filter: (h.server_id = 5)
         SubPlan 2
           ->  Result  (cost=2.25..2.26 rows=1 width=0) (actual time=0.078..0.080 rows=1 loops=1)
                 InitPlan 1 (returns $2)
                   ->  Limit  (cost=0.00..2.25 rows=1 width=8) (actual time=0.064..0.066 rows=1 loops=1)
                         ->  Index Scan Backward using tz_test_hist_rec_time_idx on tz_test_history h2  (cost=0.00..10800.38 rows=4800 width=8) (actual time=0.058..0.058 rows=1 loops=1)
                               Index Cond: ((recorded_time IS NOT NULL) AND (recorded_time < $1))
                               Filter: (server_id = $0)
         SubPlan 2
           ->  Result  (cost=2.25..2.26 rows=1 width=0) (actual time=0.078..0.080 rows=1 loops=1)
                 InitPlan 1 (returns $2)
                   ->  Limit  (cost=0.00..2.25 rows=1 width=8) (actual time=0.064..0.066 rows=1 loops=1)
                         ->  Index Scan Backward using tz_test_hist_rec_time_idx on tz_test_history h2  (cost=0.00..10800.38 rows=4800 width=8) (actual time=0.058..0.058 rows=1 loops=1)
                               Index Cond: ((recorded_time IS NOT NULL) AND (recorded_time < $1))
                               Filter: (server_id = $0)
 Total runtime: 0.525 ms
(21 rows)

--
Gurjeet Singh
EnterpriseDB Corporation
The Enterprise PostgreSQL Company

pgsql-hackers by date:

Previous
From: Noah Misch
Date:
Subject: Reducing overhead of frequent table locks
Next
From: Kevin Barnard
Date:
Subject: Re: Formatting Curmudgeons WAS: MMAP Buffers