Page 1 of 1

Hash file weired character

Posted: Wed Aug 25, 2010 9:26 am
by kab123
I have seen a weired scenario. In test environment, it writes 8 rows and reads 8 rows from hash file. but in prod, it writes 8 but reads 9 (which we think is from previous run and file was not flushed properly).. Is that possible ?? just curious to know as the job was runn ing for years and it never occured before.
'Delete file before create' not checked
'Create file' checked
'Clear file bfore creating' checked

does this cause the issue ?? because it appears it does.

Thanks

Posted: Wed Aug 25, 2010 9:41 am
by chulett
No clue how to explain that one as long as you are certain you are reading and writing to the exact same hashed file. Worst case, manually delete the hashed file's directory and it will recreate it on the next run.

Posted: Wed Aug 25, 2010 10:05 am
by kab123
100% positive on that. like i said the same job has been running differently in 2 environments. same input data.

Posted: Wed Aug 25, 2010 1:20 pm
by Sainath.Srinivasan
Write the output to a sequential file and find the difference / extra line.