Hello,
I'm getting a problem with a job that starts with a merge stage.
The Merge is on 2 text files that are created by other jobs run earlier in a sequence.
The 2 text files exist and contain the appropriate data.
The error appears in the log as follows:
Event: Starting Job PIniMedStage4.1. (...)
Event: Environment variable settings: (...)
Event: PIniMedStage4.1.SRT_o_med_date: DSD.StageRun Active stage starting, tracemode = 0.
Event: PIniMedStage4.1.MERGE_Input: Seek failed
Event: PIniMedStage4.1.MERGE_Input: Function 'put_row_to_page' failed
Event: PIniMedStage4.1.MERGE_Input: Function 'construct_hash_table' failed
Event: Attempting to Cleanup after ABORT raised in stage PIniMedStage4.1.MERGE_Input
Event: Job PIniMedStage4.1 aborted.
Event: (PIniMedSeq) <- PIniMedStage4.1: Job under control finished.
Any ideas how to address this? I can't find anything useful in the online help.
Any help will be much appreciated.
Thanks,
Louise.
Problem with Merge Stage
Moderators: chulett, rschirm, roy
Can you check the non-nullable column values, whether they contain null values. This error is causing somewhat related to NULL handling in any one of the file if my memory works correctly.
Check column metadata in the input and output whether they have same nullablity property set.
Regards
Saravanan
Check column metadata in the input and output whether they have same nullablity property set.
Regards
Saravanan
Hi,
From my experiance with this stage, If you made any changes and then had strange problems you better rebuild it from scrtch and it usually works.
IHTH,
From my experiance with this stage, If you made any changes and then had strange problems you better rebuild it from scrtch and it usually works.
IHTH,
Roy R.
Time is money but when you don't have money time is all you can afford.
Search before posting:)
Join the DataStagers team effort at:
http://www.worldcommunitygrid.org
Time is money but when you don't have money time is all you can afford.
Search before posting:)
Join the DataStagers team effort at:
http://www.worldcommunitygrid.org
Saravanan
There is one column that has null values and it is set to nullable everywhere it appears.
Roy
This is a job that was built by a contractor who was brought in to work on a number of things and has never been tested up to now, so as far as I know, it has never worked.
I tried rebuiding the merge stage anyway and have recompiled and re-run to no effect.
I'll keep working on it. I've raised a call with the helpdesk too. I'll report back if I get a resolution.
There is one column that has null values and it is set to nullable everywhere it appears.
Roy
This is a job that was built by a contractor who was brought in to work on a number of things and has never been tested up to now, so as far as I know, it has never worked.
I tried rebuiding the merge stage anyway and have recompiled and re-run to no effect.
I'll keep working on it. I've raised a call with the helpdesk too. I'll report back if I get a resolution.
According to Ascential, this is a bug in v6.0.1 which is fixed in v6.0.2 and all later releases. Can you guess which version I am working with?
I have found a workaround though. I've replaced on the of the sequential files with a hash file and am using a lookup to combine the 2 sets of data. Not as neat as using merge would have been, but at least it works!
Thanks for your help and advice.
Louise.
I have found a workaround though. I've replaced on the of the sequential files with a hash file and am using a lookup to combine the 2 sets of data. Not as neat as using merge would have been, but at least it works!
Thanks for your help and advice.
Louise.
-
- Premium Member
- Posts: 1044
- Joined: Wed Sep 29, 2004 3:30 am
- Location: Nottingham, UK
- Contact:
In our experience, the Merge stage (even in 7.0.1) is to be avoided as it brings nothing but pain. If you are doing a many-to-many join, or a full outer join, then there isn't much other choice that doesn't bring severe complexity or performance penalties though.LRenwick wrote:I have found a workaround though. I've replaced on the of the sequential files with a hash file and am using a lookup to combine the 2 sets of data. Not as neat as using merge would have been, but at least it works!
Louise.
Phil Hibbs | Capgemini
Technical Consultant
Technical Consultant