sqlite3.Connection.iterdump() does not work with row_factory = sqlite3.Row

[plemarre]

BPO	15545
Nosy	@loewis, @bitdancer
Files	iterdump.diff

^{Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.}

Show more details

GitHub fields:

assignee = None
closed_at = <Date 2013-01-10.16:35:59.047>
created_at = <Date 2012-08-03.08:07:56.234>
labels = ['type-bug', 'library']
title = 'sqlite3.Connection.iterdump() does not work with row_factory = sqlite3.Row'
updated_at = <Date 2013-01-10.16:35:58.907>
user = 'https://bugs.python.org/plemarre'

bugs.python.org fields:

activity = <Date 2013-01-10.16:35:58.907>
actor = 'r.david.murray'
assignee = 'none'
closed = True
closed_date = <Date 2013-01-10.16:35:59.047>
closer = 'r.david.murray'
components = ['Library (Lib)']
creation = <Date 2012-08-03.08:07:56.234>
creator = 'plemarre'
dependencies = []
files = ['26710']
hgrepos = []
issue_num = 15545
keywords = ['patch']
message_count = 8.0
messages = ['167296', '167297', '167543', '167579', '170875', '179559', '179566', '179568']
nosy_count = 5.0
nosy_names = ['loewis', 'peter.otten', 'r.david.murray', 'python-dev', 'plemarre']
pr_nums = []
priority = 'normal'
resolution = 'fixed'
stage = 'resolved'
status = 'closed'
superseder = None
type = 'behavior'
url = 'https://bugs.python.org/issue15545'
versions = ['Python 2.7', 'Python 3.2', 'Python 3.3', 'Python 3.4']

[loewis]

Can you please elaborate? Structure your report as follows:

1. this is what I did
2. this is what happened
3. this is what should have happened instead.

[plemarre]

I use Python 3.2.3 on GNU/Linux 64bits (openSUSE 12.2).
I have created an in-memory connection with the following code:

conn = sqlite3.connect(":memory:", detect_types=sqlite3.PARSE_DECLTYPES, check_same_thread=False)
conn.row_factory = sqlite3.Row

Then I have filled the database, but when it comes to copy it via conn.iterdump(), it crashed with the following message:

File "/usr/lib64/python3.2/sqlite3/dump.py", line 30, in _iterdump
for table_name, type, sql in sorted(schema_res.fetchall()):
TypeError: unorderable types: sqlite3.Row() < sqlite3.Row()

It seems that the error appears because of the use of "sorted()". In fact, if I don't change conn.row_factory or if I use a custom class _Row implementing __lt__ method (see below) this error does not appear.

class _Row(sqlite3.Row):
    def __lt__(self, x):
        return False

[plemarre]

By the way, this issue does not appear with Python 3.2.2.

[peterotten]

Here's a minimal fix that modifies the sql in sqlite3.dump._iterdump() to sort the tables by name. It is then no longer necessary to sort the resultset in Python for the unit tests to pass.

[plemarre]

Thanks for the patch. In which version will be your patch integrated?

[bitdancer]

For the record, this is a regression introduced by the fix for bpo-9750. I plan to commit the fix shortly, thanks for the report and patch.

[python-dev]

New changeset 2cdb599172ab by R David Murray in branch '3.2':
bpo-15545: fix sqlite3.iterdump regression on unsortable row_factory objects.
http://hg.python.org/cpython/rev/2cdb599172ab

New changeset 6a85894c428f by R David Murray in branch '3.3':
merge bpo-15545: fix sqlite3.iterdump regression on unsortable row_factory objects.
http://hg.python.org/cpython/rev/6a85894c428f

New changeset 7a62b5ee32ec by R David Murray in branch 'default':
merge bpo-15545: fix sqlite3.iterdump regression on unsortable row_factory objects.
http://hg.python.org/cpython/rev/7a62b5ee32ec

New changeset bb4e4f0cec2e by R David Murray in branch '2.7':
bpo-15545: sort iterdump via SQL instead of in python code
http://hg.python.org/cpython/rev/bb4e4f0cec2e

[bitdancer]

Peter, I see you've made contributions before, but you don't show as having a contributor agreement on file according to the tracker. Have you sent one in? If not, would you, please?

Thanks again for the fix.