Bug #81688 [Asn]: PDO_ODBC doesn't handle fixed-length character columns with character conversio

From: Date: Tue, 21 Jun 2022 13:27:28 +0000
Subject: Bug #81688 [Asn]: PDO_ODBC doesn't handle fixed-length character columns with character conversio
References: 1  Groups: php.bugs 
Request: Send a blank email to php-bugs+get-241777@lists.php.net to get a copy of this message 
Edit report at https://bugs.php.net/bug.php?id=81688&edit=1

 ID:                 81688
 Updated by:         cmb@php.net
 Reported by:        calvin at cmpct dot info
 Summary:            PDO_ODBC doesn't handle fixed-length character
                     columns with character conversio
 Status:             Assigned
 Type:               Bug
 Package:            PDO ODBC
 Operating System:   IBM i 7.2
 PHP Version:        8.0.13
 Assigned To:        cmb
 Block user comment: N
 Private report:     N

 New Comment:

Thanks for the update, and the reminder! :)

I shall have a look at this tomorrow.


Previous Comments:
------------------------------------------------------------------------
[2022-06-20 17:43:36] calvin at cmpct dot info

FWIW, as an update:

IBM released a new version of the Db2i driver that's worked around the issue for us by
reporting the size after conversion. However, the underlying bug in PDO_ODBC/odbc is still lurking
and could affect other drivers, so it's still worth pursuing a fix for it.

------------------------------------------------------------------------
[2022-01-26 20:56:02] calvin at cmpct dot info

Nitpick I have found that probably doesn't matter: odbc_result_all doesn't work with the
patch, it seems. It's deprecated, so I assume it's not a big deal, but it's something
I noticed debugging some unrelated code.

------------------------------------------------------------------------
[2022-01-20 19:11:37] calvin at cmpct dot info

Brief update w/ the PDO_ODBC patch from last week: The user does report the patch works for them
without any (noticeable) performance regression. Probably a bit rough still, but good to know so
far.

------------------------------------------------------------------------
[2022-01-14 16:49:39] cmb@php.net

> So, I've ported the logic over for PDO_ODBC and I have the patch
> here: […]

Ah, cool!  It might be possible to simplify further by just
setting S->going_long = 1 early (but that doesn't matter for now).

> coercing to string, have to get assembled in a 256 byte buffer
> piecemeal

I think that code is indeed rather suboptimal.  I still wonder why
fetching in small chunks is claimed to be faster than reading a
single chunk[1].

> PDO_ODBC is *noticeably* faster at least on Linux.

That's interesting!  Might be worth investigating.

[1] <https://github.com/php/php-src/pull/6716#issuecomment-783461335>

------------------------------------------------------------------------
[2022-01-14 16:31:56] calvin at cmpct dot info

The user having the issue wanted to try the fix with their existing code, but was using PDO. So,
I've ported the logic over for PDO_ODBC and I have the patch here: https://gist.github.com/NattyNarwhal/077dbd184f08ac030f8b6088b3d39ed4
- I'll try it out with them soon and see how it goes.

It's quite simple and like the procedural ODBC one, works by basically skipping functionality
and using existing things. There's almost certainly a better solution, as this patch does all
sorts of bad things (i.e. coercing to string, have to get assembled in a 256 byte buffer piecemeal),
but it does fix the primary issue here.

(Also, unrelated observation and happens with/without any of the patches: PDO_ODBC is *noticeably*
faster at least on Linux. I can actually see the results stream out w/ procedural, but they're
instant with PDO, and that's without fetchAll. Quite interesting...)

------------------------------------------------------------------------


The remainder of the comments for this report are too long. To view
the rest of the comments, please view the bug report online at

    https://bugs.php.net/bug.php?id=81688


--
Edit this bug report at https://bugs.php.net/bug.php?id=81688&edit=1

Thread (37 messages)

« previous php.bugs (#241777) next »