db-derby-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Mike Matrigali <mikem_...@sbcglobal.net>
Subject Re: New predicatePushdown.sql diff...
Date Mon, 06 Mar 2006 17:14:41 GMT
I agree that best solution is to change the ij output to have the entire
plan printed.

The usual cause for this is not a difference in when ij truncates, it
is an actual change in the output.  It is usually not obvious from the
diff because we diff a "sed'd" output in the case of query plans and
the difference is usually in the the stuff sed'd out - like estimated
rows (ie. going from 1 char to 2 chars) which may differ depending
on timing.  Or it could be a difference on that jvm on how it formats 
some datatype.

I don't know if that is the case here, just what I have seen in the past.

Army wrote:
> Latest tinderbox results are showing a diff for a test I added as part 
> of DERBY-805:
> *** Start: predicatePushdown jdk1.5.0_04 derbyall:derbylang 2006-03-04 
> 00:47:27 ***
> 6065 del
> <                                     Number of row&
> 6065a6065
>  >                                     Number of rows visi&
> Test Failed.
> I ran the test on Windows 2000 using IBM 1.4.2, Sun 1.4.2, and Sun 1.5 
> and did not see this diff.  I also ran on Linux Red Hat using IBM 1.4.2 
> and didn't see the diff.
> It looks like the row returned from ij is truncated a few characters 
> later than expected--but I don't know what could be causing that...?
> In any event, I'll try to increase the length of max width of the column 
> to avoid truncation; hopefully that will get rid of the diff.
> Army

View raw message