Opened 13 years ago
Closed 12 years ago
#17158 closed Bug (fixed)
django.db.backends.BaseDatabaseOperations.last_executed_query can raise an exception
Reported by: | Owned by: | Aymeric Augustin | |
---|---|---|---|
Component: | Database layer (models, ORM) | Version: | 1.2 |
Severity: | Normal | Keywords: | |
Cc: | rmanoch@… | Triage Stage: | Ready for checkin |
Has patch: | yes | Needs documentation: | no |
Needs tests: | no | Patch needs improvement: | no |
Easy pickings: | no | UI/UX: | no |
Description
Method "last_executed_query" should not always return "smart_unicode(sql) % u_params". If params variable is empty it should return just "smart_unicode(sql)". The problem is when we try to execute query with MySQL DATA_FORMAT function it says error.
Example:
from django.db import connection c = connection.cursor() c.execute("SELECT * FROM VAR_current GROUP BY DATE_FORMAT(closing_month, '%Y-%m');") print c.fetchall()
Error:
django/db/backends/__init__.py", line 216, in last_executed_query return smart_unicode(sql) % u_params TypeError: not enough arguments for format string
This issue is closely related to #9055.
Attachments (2)
Change History (8)
comment:1 by , 13 years ago
Summary: | django.db.backends.BaseDatabaseOperations.last_executed_query issue → django.db.backends.BaseDatabaseOperations.last_executed_query can raise an exception |
---|---|
Triage Stage: | Unreviewed → Accepted |
by , 13 years ago
Attachment: | 17158-testcase.patch added |
---|
comment:2 by , 13 years ago
I've just uploaded a test case demonstrating the issue under SQLite. The assertion in this test case may have to be changed depending on the solution that we actually choose.
comment:3 by , 12 years ago
Owner: | changed from | to
---|
by , 12 years ago
Attachment: | 17158.patch added |
---|
comment:4 by , 12 years ago
Has patch: | set |
---|---|
Triage Stage: | Accepted → Ready for checkin |
I'll commit the patch I just attached if no one objects.
comment:6 by , 12 years ago
Resolution: | → fixed |
---|---|
Status: | new → closed |
Since r16081, there is a better implementation of
last_executed_query
for MySQL (and Oracle), so the problem is solved as far as MySQL is concerned.It still exists for SQLite, and although we came to the conclusion that we couldn't build a reliable
last_executed_query
for SQLite in #14091, it shouldn't raise an exception.I see three possible solutions:
I'm in favor of 1 because it will avoid returning subtly invalid SQL when the database backend in SQLite. It's backwards incompatible, but
last_executed_query
never returned something usable under SQLite. Returning obviously incomplete SQL is less likely to confuse newcomers.I'm against 3 because it's overengineering, and I'm not convinced it's possible to get it right for all cases anyway.