public inbox for [email protected]
help / color / mirror / Atom feedFrom: Neto pr <[email protected]>
To: Igor Neyman <[email protected]>
Cc: [email protected] <[email protected]>
Subject: Re: Explain Analyze - actual time in loops
Date: Fri, 8 Sep 2017 06:08:08 -0700
Message-ID: <CA+wPC0ObeUigZk3Vf_EVtozKUoCcd91FDFUd4VR4L_3LkY6bbA@mail.gmail.com> (raw)
In-Reply-To: <DM5PR17MB14993F6FD18158C593907D85DA950@DM5PR17MB1499.namprd17.prod.outlook.com>
References: <CA+wPC0OXdN+_8HxhYMWyV9nfD8v+xChvLROiWCiRo=qSve2MjQ@mail.gmail.com>
<DM5PR17MB14993F6FD18158C593907D85DA950@DM5PR17MB1499.namprd17.prod.outlook.com>
List-Unsubscribe: <mailto:[email protected]?body=unsub%20pgsql-performance>
Igor,
You're right, I confused the radix character.
But even so the result is approximate to the previous message, 182 minutes, see
below:
419.113 / 1000 = 0.41 seconds * 26469 (loops) = 11093.50 seconds or 184
minutes
After analyzing, I saw that in some places of the plan, it is being used
Parallelism. Does this explain why the final value spent (in minutes) to go
through the index (184 minutes) is greater than the total query time (66
minutes)?
Regards
Neto
2017-09-08 5:46 GMT-07:00 Igor Neyman <[email protected]>:
> *From:* [email protected] [mailto:pgsql-performance-
> [email protected]] *On Behalf Of *Neto pr
> *Sent:* Thursday, September 07, 2017 11:17 PM
> *To:* [email protected]
> *Subject:* [PERFORM] Explain Analyze - actual time in loops
>
>
>
> …
>
> ################################ ###################################
> -> Index Scan using idx_l_partkeylineitem000x on lineitem (cost =
> 0.57..97.65 rows = 26 width = 36)
> (current time = 23.615..419.113 rows = 30 loops = 26469)
> Index Cond: (l_partkey = part.p_partkey)
> ################################################## #################
> According to the documentation, one should multiply the Actual Time by the
> number of Loops.
> That is: 419113 ms -> 419113/1000/60 = 6.9 minutes * 26469 (loops) = 182.6
> minutes.
>
> But how does this stretch take 182.6 minutes, if the entire query ran in
> 66 minutes?
>
> ……………….
> thank you and best regards
> [] 's Neto
>
> Neto,
>
> The time you see there is in ms, so the point (‘.’) you see is the digital
> point.
>
> So, it is 419.113ms or a little less than half a second (0.419sec).
>
> Igor Neyman
>
reply
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Reply to all the recipients using the --to and --cc options:
reply via email
To: [email protected]
Cc: [email protected], [email protected]
Subject: Re: Explain Analyze - actual time in loops
In-Reply-To: <CA+wPC0ObeUigZk3Vf_EVtozKUoCcd91FDFUd4VR4L_3LkY6bbA@mail.gmail.com>
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
This inbox is served by agora; see mirroring instructions
for how to clone and mirror all data and code used for this inbox