#273 lazyUpdate is resetting fields to default values when read

closed-works-for-me
General (125)
5
2013-04-28
2012-09-13
Anonymous
No

I'm trying to use SQLObject's lazyUpdate feature and my data is getting clobbered. Here's the class definition

class T(sqlobject.SQLObject):
class sqlmeta:
lazyUpdate = True
def __enter__(self):
return self
def __exit__(self, type, value, traceback):
if self.sqlmeta.dirty:
self.syncUpdate()
return False
foo = IntCol()
bar = IntCol(default=None)

The context manager functions are so I can write something like:

with T.select().getOne() as t:
# update t

Everything seems to work fine when I'm creating the objects. However, when I'm just reading the objects, fields with default values (like bar above) are getting reset to their default values.

For example, given

def get_rows():
rows = T.select()
return list(repr(r) for r in rows)

If I call get_rows repeatedly, I find that the first time I get the correct data and subsequent times some or all of the rows have their values replaced by their defaults. Eventually all the rows are reset to defaults. Fields without default values are not touched.

If I comment out the line that sets lazyUpdate in the class definition, it works fine (presumably with many more writes to the db). If I comment out the call to syncUpdate it still fails. Changing the default to a different value (like -1) still fails (and values get reset to -1 not None). It's not some lingering schema setting as I've deleted the table and recreated it multiple times as I've been trying to resolve the problem.

FWIW, I'm using SQLObject 1.3.1, Python 2.7 and PostreSQL 9.1.

Discussion

  • Oleg Broytman

    Oleg Broytman - 2012-09-21
    • assigned_to: nobody --> phd
    • status: open --> open-works-for-me
     
  • Oleg Broytman

    Oleg Broytman - 2012-09-21

    I cannot reproduce the problem; SQLObject 1.3.0, SQLite, Postgres. Can you write a simple test scripts (or a pair of scripts -- an initialization and a test) that demonstrates the problem? As simple as possible, i.e. without making the table a context manager and so on.

     
  • Comment has been marked as spam. 
    Undo

    You can see all pending comments posted by this user  here

    Anonymous - 2012-10-03

    This resolution is not helpful. It doesn't work for me. This is a catastrophic failure. It doesn't reset the data on every read but it always does eventually.

    Without a fix, I have no choice but to stop using sqlobject.

     
  • Oleg Broytman

    Oleg Broytman - 2012-10-03

    And without reliable way to reproduce the problem I cannot provide a fix. I
    cannot even be sure it's a problem in SQLObject and not in your program or
    other libraries. Please demonstrate the bug.

     
  • Comment has been marked as spam. 
    Undo

    You can see all pending comments posted by this user  here

    Anonymous - 2012-10-03

    Sorry about that. I've never used a bug tracking system before that hides the conversation so I didn't see your comment. All I saw was the resolution change. It repros very reliably in my program. I can't give you my whole program but I can try and pare it down.

     
  • Oleg Broytman

    Oleg Broytman - 2012-10-03

    Yes, SF tracker is not the best tool for beginners; it takes some time to master it. The smaller script you can provide the better. Waiting...

     
  • Oleg Broytman

    Oleg Broytman - 2012-10-16

    Any progress with test scripts?

     
  • Oleg Broytman

    Oleg Broytman - 2013-04-28
    • status: open-works-for-me --> closed-works-for-me
     

Get latest updates about Open Source Projects, Conferences and News.

Sign up for the SourceForge newsletter:





No, thanks