Optimizing Update with WHERE x IN (id1,id2,...) - Mailing list pgsql-admin

Dear devs,

I just saw this EXPLAIN in PostgreSQL 8.3.12:

EXPLAIN UPDATE dbmail_messages SET recent_flag = 0 WHERE message_idnr IN
(3506054,3506052,3506050,3506048,3506046,3506044,3506042,3506040,3506038,3506036,3506034,3506032,3506030)
AND recent_flag = 1;
                                                                            QUERY
PLAN

------------------------------------------------------------------------------------------------------------------------------------------------------------------
 Bitmap Heap Scan on dbmail_messages  (cost=55.39..105.04 rows=1
width=75)
   Recheck Cond: (message_idnr = ANY

('{3506054,3506052,3506050,3506048,3506046,3506044,3506042,3506040,3506038,3506036,3506034,3506032,3506030}'::bigint[]))
   Filter: (recent_flag = 1)
   ->  Bitmap Index Scan on dbmail_messages_pkey  (cost=0.00..55.39
rows=13 width=0)
         Index Cond: (message_idnr = ANY

('{3506054,3506052,3506050,3506048,3506046,3506044,3506042,3506040,3506038,3506036,3506034,3506032,3506030}'::bigint[]))
(5 rows)

And it seems to me that could be optimized to skip the "Recheck Cond" as
that is already checked by the Index Scan?  Should save some CPU cycles.

--
mit freundlichen Grüssen,
Michael Monnerie, Ing. BSc

it-management Internet Services: Protéger
http://proteger.at [gesprochen: Prot-e-schee]
Tel: +43 660 / 415 6531

// ****** Radiointerview zum Thema Spam ******
// http://www.it-podcast.at/archiv.html#podcast-100716
//
// Haus zu verkaufen: http://zmi.at/langegg/