db-derby-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Rami Ojares / PDF-Comics Oy <r...@absinth.fi>
Subject Re: delete from table
Date Tue, 12 Jan 2010 11:21:34 GMT
Thank you all for help and answers.

I can't use the import procedure's replace parameter, because I import 
from another db a and do elaborate transformations to the data on the way.

Dropping tables is faster.
I have this sql file where I drop all triggers, views and tables in 
proper order to drop the whole schema.
But I always have to run it twice after an import.
At some point (I am not sure if it changes or is always the same) the 
execution of those sql statements hangs for some time
and then finally returns with an exception:
"A lock could not be obtained within the time requested".
Then when I start it again at the point where the exception occurred 
everything goes fine (and fast).

Even with this one complication it is much faster to drop the schema 
than do DELETE FROM TBL for 5 tables in the schema.
(One of those tables has over 10 000 images in them but I don't think 
that's causing the problems any more than the other tables.)

I won't start using debug version just because I wish to minimize all 
the admin hassles.
And I wouldn't know where to look for the syntax (and I won't go digging 
source for a problem like this).
But I voted for issue DERBY-268.
And JIRA said it has been approved as part of sql standard.
Although I don't really know why derby community holds compliancy in 
such a high regard.
None of the databases I have ever worked with are really - in practice - 
But I have to admit that IBM databases (DB2 and Derby) are the most 
meticulously documented.
And that's a big plus.
And maybe compliancy keeps derby somehow defined and brings clear 
guidelines to what can be included and what not.
Oh, but I digress.

What I really meant to say was thank you for the advice.

- rami

View raw message