Debugging SQL queries

Anastasios Hatzis <ahatzis@gmx.net> writes:

My problem is that I don't know where the given character number is
inside the query/file. I can search in my code editors (e.g. those of
Eclipse IDE) for line number + character number in given line
number... but I can not search for an overall character
number.

Emacs has the "goto-char" function and runs on Windows, so you could
use that (there are probably other editors with this feature as
well)..

-Doug

Douglas McNaught wrote:

Emacs has the "goto-char" function and runs on Windows, so you could
use that (there are probably other editors with this feature as
well)..

Doug, thank you. I will use it. - Ha, I never thought I will ever use
legendary Emacs. Probably I will even become an entry-class geek now? ;-)

Anastasios

From my point of view, more important problem is that log doesn't help
to find the query (in other words, log message doesn't show context)

On 5/11/06, Douglas McNaught <doug@mcnaught.org> wrote:

Anastasios Hatzis <ahatzis@gmx.net> writes:

My problem is that I don't know where the given character number is
inside the query/file. I can search in my code editors (e.g. those of
Eclipse IDE) for line number + character number in given line
number... but I can not search for an overall character
number.

Emacs has the "goto-char" function and runs on Windows, so you could
use that (there are probably other editors with this feature as
well)..

-Doug

---------------------------(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

--
Best regards,
Nikolay

On 5/11/06, Nikolay Samokhvalov <samokhvalov@gmail.com> wrote:

From my point of view, more important problem is that log doesn't help
to find the query (in other words, log message doesn't show context)

Yes it does.. but it depends on your logging setup.

in psql:

test=# blah;
ERROR: syntax error at or near "blah" at character 1

my log shows:

... [9-1] LOG: statement: blah;
... [10-1] ERROR: syntax error at or near "blah" at character 1

Do you have:

log_statement = true

in your postgresql.conf ?

--
Postgresql & php tutorials
http://www.designmagick.com/

Anastasios Hatzis <ahatzis@gmx.net> writes:

I have sometimes very long queries, for example in setup-purposed sql
file where plenty of tables of a new database are created in a single
transaction. 600 lines of code or more is common. Sometimes, while
developing, an error may appear after submitting such a query for
testing purposes via pgAdmin III query dialog. pgAdmin III displays the
PostgreSQL error, like shown below (example)

syntax error at or near "1" at character 2641

My problem is that I don't know where the given character number is
inside the query/file.

You should complain to the pgAdmin people that they are delivering an
inadequate error message. psql, for example, does something like this
in similar situations:

regression=# select 1 + 1 +1
regression-# + 1 + 1 1 + 1
regression-# + 1 + 1;
ERROR: syntax error at or near "1" at character 26
LINE 2: + 1 + 1 1 + 1
^
regression=#

If you ask me, a GUI that fails to make use of the cursor-position info
that the backend supplies is a poor excuse for a GUI. We have actually
removed the "at character N" bit from the default message format in CVS
tip, on the assumption that every client app worth worrying about has
got better methods than that for showing the error cursor position.
pgAdmin needs to catch up before 8.2 comes out ;-)

regards, tom lane

-----Original Message-----
From: pgsql-general-owner@postgresql.org
[mailto:pgsql-general-owner@postgresql.org] On Behalf Of Tom Lane
Sent: 11 May 2006 15:00
To: Anastasios Hatzis
Cc: pgsql-general@postgresql.org; pgadmin-hackers@postgresql.org
Subject: Re: [GENERAL] Debugging SQL queries

pgAdmin needs to catch up before 8.2 comes out ;-)

pgAdmin marks the line that the failure occurred at, and has done for
years. Whether or not it should dump the cursor at the exact location is
another issue altogether - it might upset ongoing editting (the user
could be editting or adding to the query whilst a long running part is
already running).

Regards, Dave.

Tom Lane wrote:

We have actually removed the "at character N" bit from the default message format in CVS

Tom,
What will the error format be for 8.2? Since the char number has been
removed, what will it show?
Line number?

Thanks,

--
Tony Caduto
AM Software Design
http://www.amsoftwaredesign.com
Home of PG Lightning Admin for Postgresql
Your best bet for Postgresql Administration

Tony Caduto <tony_caduto@amsoftwaredesign.com> writes:

Tom Lane wrote:

We have actually removed the "at character N" bit from the default message format in CVS

What will the error format be for 8.2?

In CVS tip you just see

regression=# select 1 2;
ERROR: syntax error at or near "2"
LINE 1: select 1 2;
^
regression=#

We still include "at character N" if the message is written to the
server log, but the "primary message text" field of the error report
sent to the client doesn't have it any more. This could still be
revisited if anyone's really unhappy about it, but ISTM that if the
client app is holding up its end about using the error-position field,
putting it in the text is just clutter.

regards, tom lane

In CVS tip you just see

regression=# select 1 2;
ERROR: syntax error at or near "2"
LINE 1: select 1 2;
^
regression=#

We still include "at character N" if the message is written to the
server log, but the "primary message text" field of the error report
sent to the client doesn't have it any more. This could still be
revisited if anyone's really unhappy about it, but ISTM that if the
client app is holding up its end about using the error-position field,
putting it in the text is just clutter.

regards, tom lane

Hi Tom,
I guess I didn't know there was a error-position field returned, I have
been parsing out the "at character xx" to get the char number.

Do you know offhand if there is the error-position field is discussed in
the docs anywhere?

Thanks,

--
Tony Caduto
AM Software Design
http://www.amsoftwaredesign.com
Home of PG Lightning Admin for Postgresql
Your best bet for Postgresql Administration

On Thu, May 11, 2006 at 10:56:20AM -0400, Tom Lane wrote:

We still include "at character N" if the message is written to the
server log, but the "primary message text" field of the error report
sent to the client doesn't have it any more.

Will the text returned by PQerrorMessage still include the character
position?

Wayne Conrad

Hi Tom,
I guess I didn't know there was a error-position field returned, I have
been parsing out the "at character xx" to get the char number.

Do you know offhand if there is the error-position field is discussed in
the docs anywhere?

Not a C guy but possibly?

http://www.commandprompt.com/community/pgdocs81/libpq-exec
http://www.commandprompt.com/community/pgdocs81/protocol-error-fields

Sincerely,

Joshua D. Drake
--

=== The PostgreSQL Company: Command Prompt, Inc. ===
Sales/Support: +1.503.667.4564 || 24x7/Emergency: +1.800.492.2240
Providing the most comprehensive PostgreSQL solutions since 1997
http://www.commandprompt.com/

Tony Caduto <tony_caduto@amsoftwaredesign.com> writes:

I guess I didn't know there was a error-position field returned, I have
been parsing out the "at character xx" to get the char number.

Do you know offhand if there is the error-position field is discussed in
the docs anywhere?

See PQresultErrorField().

regards, tom lane

On Thu, May 11, 2006 at 09:39:59AM -0500, Tony Caduto wrote:

Tom Lane wrote:

We have actually removed the "at character N" bit from the default message
format in CVS

Tom,
What will the error format be for 8.2? Since the char number has been
removed, what will it show?
Line number?

bench=# SELECT blah
bench-# FROMx;
ERROR: parse error at or near "FROMx"
LINE 2: FROMx;
^
bench=#

Note that it's still there in the server logs, though...

ERROR: parse error at or near "FROMx" at character 13
--
Jim C. Nasby, Sr. Engineering Consultant jnasby@pervasive.com
Pervasive Software http://pervasive.com work: 512-231-6117
vcard: http://jim.nasby.net/pervasive.vcf cell: 512-569-9461