From: | Chris Travers <chris(dot)travers(at)gmail(dot)com> |
---|---|
To: | Jeff Adams <Jeff(dot)Adams(at)noaa(dot)gov> |
Cc: | pgsql-general(at)postgresql(dot)org |
Subject: | Re: SQL Help - Finding Next Lowest Value of Current Row Value |
Date: | 2011-10-01 22:50:26 |
Message-ID: | CAKt_Zfs-FE9m=hgPwgS2mLXOC117NaWrMkrNZkfE8adxkKnryw@mail.gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
On Sat, Oct 1, 2011 at 2:30 PM, Jeff Adams <Jeff(dot)Adams(at)noaa(dot)gov> wrote:
> Greetings,
>
> I have a large table (~19 million records). Records contains a field
> identifying a vessel and a field containing an time (epoch). Using the
> current rows vessel and time values, I need to be able to find the next
> lowest time value for the vessel and use it to compute how much time has
> elapsed between the records. I have tried a scalar subquery in the SELECT,
> which works, but it runs quite slowly. Is there an alternative approach that
> might perform better for this type of query. Any information would be
> greatly appreciated. Thanks...
>
How many rows per vessel?
Either way it seems to me we are talking about selecting the current
record, and then selecting the max of another record within a
timeframe. I would probably use a plpgsql function that would do this
in two stages, ideally being able to do an index scan twice....
Best Wishes,
Chris Travers
From | Date | Subject | |
---|---|---|---|
Next Message | Reuven M. Lerner | 2011-10-01 23:25:03 | Searching for "bare" letters |
Previous Message | Jeff Adams | 2011-10-01 21:30:47 | SQL Help - Finding Next Lowest Value of Current Row Value |