Postgres Pro
Downloads
Home   >   mailing lists

Re: Bug in nbtree optimization to skip > operator comparisons (or < comparisons in backwards scans) - Mailing list pgsql-hackers

From Pavel Borisov
Subject Re: Bug in nbtree optimization to skip > operator comparisons (or < comparisons in backwards scans)
Date
Msg-id CALT9ZEFRXvd9L5qbqYkBEBvJi=CYXpNPt0Bcs5eKgw06=fM17w@mail.gmail.com
Whole thread Raw
In response to Re: Bug in nbtree optimization to skip > operator comparisons (or < comparisons in backwards scans)  ("Anton A. Melnikov" <a.melnikov@postgrespro.ru>)
Responses Re: Bug in nbtree optimization to skip > operator comparisons (or < comparisons in backwards scans)
List pgsql-hackers
Tree view
Hi, Anton!
Looks like an oversight when refactoring BTScanOpaqueData.firstPage into using function argument in 06b10f80ba4.

@@ -2487,14 +2486,13 @@ _bt_endpoint(IndexScanDesc scan, ScanDirection dir)
104
105     /* remember which buffer we have pinned */
106     so->currPos.buf = buf;
107 -   so->firstPage = true;
108
109     _bt_initialize_more_data(so, dir);
110
111     /*
112      * Now load data from the first page of the scan.
113      */
114 -   if (!_bt_readpage(scan, dir, start))
115 +   if (!_bt_readpage(scan, dir, start, false))

Attached is a fix.
Thank you!

Regards,
Pavel


On Fri, 22 Mar 2024 at 11:29, Anton A. Melnikov <a.melnikov@postgrespro.ru> wrote:
Hi!

Maybe _bt_readpage(scan, dir, start, true) needed at this line:

https://github.com/postgres/postgres/blob/b4080fa3dcf6c6359e542169e0e81a0662c53ba8/src/backend/access/nbtree/nbtsearch.c#L2501

?

Do we really need to try prechecking the continuescan flag here?

And the current "false" in the last arg does not match the previous code before 06b10f80ba
and the current comment above.

Would be very grateful for clarification.

With the best regards!

--
Anton A. Melnikov
Postgres Professional: http://www.postgrespro.com
The Russian Postgres Company
Attachment

pgsql-hackers by date:

Previous
From: "Anton A. Melnikov"
Date:
Subject: Re: Bug in nbtree optimization to skip > operator comparisons (or < comparisons in backwards scans)
Next
From: "Anton A. Melnikov"
Date:
Subject: Re: Bug in nbtree optimization to skip > operator comparisons (or < comparisons in backwards scans)