Re: Query plan and execution time of a query - Mailing list pgsql-hackers

From Kenneth Marshall
Subject Re: Query plan and execution time of a query
Date
Msg-id 20070721201049.GN26122@it.is.rice.edu
Whole thread Raw
In response to Query plan and execution time of a query  ("Mickael DELOISON" <mdeloison@gmail.com>)
Responses Re: Query plan and execution time of a query  (Andrew Dunstan <andrew@dunslane.net>)
List pgsql-hackers
Mickael,

Not knowing your query, it sounds like your method is working
correctly. It is quite normal to have the initial query take
longer than subsequent queries. This is a cache effect and is
what databases, in general, strive for performance-wise. I
suspect that the second time you run the query that it does
not take 0ms, just less than 1ms.

Ken

On Sat, Jul 21, 2007 at 03:26:19PM +0200, Mickael DELOISON wrote:
> Hello,
> 
> I am trying to get the execution time of a query by using JDBC (Java).
> To do that I make the difference between the time (in milliseconds)
> after the query execution and the time before the query execution.
> 
> The first time I do that, I get the real execution time, but the query
> plan seems to be cached. And the next time I do the same operation,
> for a simple query I do not get the real time (in fact  my 'simple'
> query seems to execute in 0 millisecond).
> 
> So I need help: I would like to solve this problem but I do not know
> how. I have tried some researches on Internet but I did not find. Is
> there a possibility to disable the query plan cache temporarily? It is
> also possible I am totally wrong and the problem is not where I see
> it... Anyway I need help.
> 
> Thank you in advance,
> Mickael
> 
> ---------------------------(end of broadcast)---------------------------
> TIP 9: In versions below 8.0, the planner will ignore your desire to
>       choose an index scan if your joining column's datatypes do not
>       match
> 


pgsql-hackers by date:

Previous
From: "Pavel Stehule"
Date:
Subject: Re: 8.2 is 30% better in pgbench than 8.3
Next
From: Andrew Dunstan
Date:
Subject: Re: Query plan and execution time of a query