Fatal Error: Field not found: keyChange

rajan.n · Post by **rajan.n** » Wed Jul 16, 2008 12:37 pm

Hi

While using the Reference Match stage in my job , The job aborts throwing the error

"main_program: Fatal Error: Field not found: keyChange"

The prev message of this fatal one is

Reference_Match_12: Creating sub-operator: <QSmatStats -workDir ./RT_QS37/V0S12 -f Reference_Match_12.txt>
Creating sub-operator: <copy>
Creating sub-operator: <generator -par -schema "record(qsMatchDataID:uint64{cycle={init='part+1',incr=partcount}})">
Creating sub-operator: <copy>
Creating sub-operator: <copy>
Creating sub-operator: <generator -par -schema "record(qsMatchDataID:uint64{cycle={init='part+1',incr=partcount}})">
Creating sub-operator: <copy>
Creating sub-operator: <QSmat -n 1 -workDir ./RT_QS37/V0S12 -f Reference_Match_12.txt>
Creating sub-operator: <tsort -key qsMatchDataIDA -collation_sequence en -strength secondary>
Creating sub-operator: <QSmat_join -workDir ./RT_QS37/V0S12 -f Reference_Match_12.txt>
Creating sub-operator: <tsort -key qsMatchDataIDB -collation_sequence en -strength secondary>
Creating sub-operator: <QSmat_join -B -workDir ./RT_QS37/V0S12 -f Reference_Match_12.txt>
Creating sub-operator: <switch -ignoreRejects -key qsMatchType -case MP=0 -case CP=1 -case DA=2 -case DB=3>

I have already see a post on this :

viewtopic.php?t=114578&highlight=Fatal+ ... +keyChange

I have also given one blocking coloumn still the error is not solved.

can you please share the solutioin if any one has comeacross same problem?

Thanks,
Rajan

emeri1md · Post by **emeri1md** » Wed Jul 16, 2008 12:56 pm

I replied to the previous post as well (the last entry).

Make sure you are blocking on something in all of the passes in the Match Spec. I made that mistake as well, and spent hours trying to fix it.

rajan.n · Post by **rajan.n** » Wed Jul 16, 2008 1:05 pm

Hi Emeri,

I have tried in that way too.

I have included 5 Match commands
say :
src_name - ref_name
src_addr - ref_addr
src_city - ref_city
src_stage - ref_state
src_zip - ref_zip
, and for the same coloums i have included the Blocking columns.

still the error exists.....

Can you please let me know the blocking colums need to be given any criteria ?

emeri1md wrote:I replied to the previous post as well (the last entry).

Make sure you are blocking on something in all of the passes in the Match Spec. I made that mistake as well, and spent hours trying to fix it.

emeri1md · Post by **emeri1md** » Wed Jul 16, 2008 1:34 pm

Not to sound condescending, but you did provision the match spec and passes, and compile the job, right? I've made that mistake myself.

All blocking does is group the data according to the values. This makes the actual matching better. You should try to use the standardized data, probably some of the NYSIS entries. Maybe just block on the City NYSIS and the State for starters.

rajan.n · Post by **rajan.n** » Wed Jul 16, 2008 2:18 pm

Yes I did provision all...!!

I wanted to do a basic comparision. using the original columns.( Not using any of standerdize generated columns). But still i see the same error..

one more question : when we give the test parameters in the match specification do i need to give the DB name DSN name there..? does it need only for testing code ? or its maditary to give all the values to run a job ?

emeri1md · Post by **emeri1md** » Wed Jul 16, 2008 2:29 pm

Configuring the DB connection is only for testing the pass within the Match Spec. You don't need it to run the job itself. If you're able to set up a test DB for testing the passes, I'd go for it; it seems to be faster.

I'm not sure how much more help I'll be for your problem. I've been dealing with a lot of other problems since, so the solution, other than blocking, slips me mind. Sorry.

rajan.n · Post by **rajan.n** » Wed Jul 16, 2008 2:40 pm

Thank you so much for your time, still that was a learning stuff.

Can you please give the process of performing the following task ? at least meight help me wether am on the path....

I have 2 source one form database and other form a file. the names are organizations name , other are some address fields.
Want to comapare those 2 source to find the matches / and unmatches.
I want to try on Name first as the basic start.

Thanks again for your time.

emeri1md · Post by **emeri1md** » Thu Jul 17, 2008 6:08 am

I wouldn't recommend matching on only the name. We have a group where that's all we have, and our success rate is low. Always try to match on address if you can.

And the area info (city/state/zip) are good for blocking.

Good luck.

rajan.n · Post by **rajan.n** » Thu Jul 17, 2008 10:27 am

Thank you ,

All I did is created again the standardise files and frequency files for both source and ref data. then i configured the match specification again with only Name and blocking with city ( source and ref).

Job ran fine

Thanks emeri1md once again.

emeri1md wrote:I wouldn't recommend matching on only the name. We have a group where that's all we have, and our success rate is low. Always try to match on address if you can.

And the area info (city/state/zip) are good for blocking.

Good luck.