BUG #6271: psql -f reporting unexpected syntax errors on first command
The following bug has been logged online:
Bug reference: 6271
Logged by: Shannon
Email address: shannon@arc.net.au
PostgreSQL version: 8.4.9
Operating system: Gentoo Linux 64bit
Description: psql -f reporting unexpected syntax errors on first
command
Details:
Given the following 2 line SQL script in a file:
--- test.sql ---
SHOW client_encoding;
SHOW client_encoding;
----------------$ psql -ftest.sql -U postgres template1
psql:test.sql:1: ERROR: syntax error at or near "SHOW"
LINE 1: SHOW client_encoding;
^
client_encoding
-----------------
UTF8
(1 row)
See what happened there? The first line is treated as a syntax error and the
second line works which is insane because the two lines are identical!
I've tried:
* Adding empty lines and/or comments to the top of the file
* Setting PGCLIENTENCODING in the shell to a variety of values including
UTF8, SQL_ASCII and LATIN1
* Other commands
What works:
* Every line except the one with the first command
* Piping a string into psql stdin
* Using psql interactively
Other factors:
Could this be related to my locale?
$ locale
LANG=en_AU.utf8
LC_CTYPE="en_AU.utf8"
LC_NUMERIC="en_AU.utf8"
LC_TIME="en_AU.utf8"
LC_COLLATE="en_AU.utf8"
LC_MONETARY="en_AU.utf8"
LC_MESSAGES="en_AU.utf8"
LC_PAPER="en_AU.utf8"
LC_NAME="en_AU.utf8"
LC_ADDRESS="en_AU.utf8"
LC_TELEPHONE="en_AU.utf8"
LC_MEASUREMENT="en_AU.utf8"
LC_IDENTIFICATION="en_AU.utf8"
LC_ALL=
On ons, 2011-10-26 at 04:08 +0000, Shannon wrote:
The following bug has been logged online:
Bug reference: 6271
Logged by: Shannon
Email address: shannon@arc.net.au
PostgreSQL version: 8.4.9
Operating system: Gentoo Linux 64bit
Description: psql -f reporting unexpected syntax errors on first
command
Details:Given the following 2 line SQL script in a file:
--- test.sql --- SHOW client_encoding; SHOW client_encoding; ----------------$ psql -ftest.sql -U postgres template1
psql:test.sql:1: ERROR: syntax error at or near "SHOW"
LINE 1: SHOW client_encoding;
^
client_encoding
-----------------
UTF8
(1 row)
Look at the file with a hex editor or od or something like that to see
whether there are any "invisible" characters on the first line. A
Unicode byte-order mark is a possibility.
Excerpts from Shannon's message of mié oct 26 01:08:00 -0300 2011:
Given the following 2 line SQL script in a file:
--- test.sql --- SHOW client_encoding; SHOW client_encoding; ----------------$ psql -ftest.sql -U postgres template1
psql:test.sql:1: ERROR: syntax error at or near "SHOW"
LINE 1: SHOW client_encoding;
^
client_encoding
-----------------
UTF8
(1 row)See what happened there? The first line is treated as a syntax error and the
second line works which is insane because the two lines are identical!
They are not identical -- the first line contains a BOM, which I
actually see in the error messages you pasted (but you likely don't):
psql:test.sql:1: ERROR: syntax error at or near "<feff>SHOW"
LINE 1: <feff>SHOW client_encoding;
This is being hidden from you by your text editor. You need to remove
those bytes before passing the file to psql.
--
Álvaro Herrera <alvherre@commandprompt.com>
The PostgreSQL Company - Command Prompt, Inc.
PostgreSQL Replication, Consulting, Custom Development, 24x7 support