change capture

[knowledge]

Hi ,

I am using scd type 2 ,

I am using change capture stage , which is working fine with the sample data , but when I am trying to load the real data , update is failing , though the records in before and after sets are similar it is giving update that is change_code=3 ,

I ceated two table with 5 records which are exact copy , then from after data set I changed 2 for update , 1 for insert and 1 delete , I got the exact result , but with real data total count for after data(source) set 414925 , before (target ) data set is 414926 , output of change capture stage gives me 414928 out of which 414923 are update 3 delete and 2 insert ,

but when I checked the update data it is exactly same as target source ,
it shouldnt give update it should be copy ,

I selected the option implicit keys and all values and I did not even include the field like rec_effective_date, rec_exeration_date or delete_flag which may cause the problem , but then also I am not getting exact count for update ,

please suggest ,

[vmcburney]

Be very careful with CHAR and VARCHAR combinations and trimming as sometimes your before values get implicitely modified and no longer match your after values. Try removing columns from the value list one at a time via explicit value settings until you find which field is causing the problem.

[knowledge]

Thanks for the replay ,

But I have used the option explicite keys and all values , and the before and after meta data table definition are loaded from the same that is target table meta data ,

then also the meta data will cause the problem ,

pl suggets ,

thanks

[kumar_s]

HI,
If that is the case it shouldn't be a problem.
But let us know how you determine that the both files are identical. For testing purpose just make a copy of the same file and give the two inputs and check.
It shoudl work fine.

-Kumar

[thebird]

Hi,

Also make sure that you have used the right partitioning mode for the Change capture stage. I believe that incorrect modes can also cause problems like this. Suggest that you use the Hash partitioning on the After and Before datasets .

Regards,

The Bird.

[knowledge]

Hi ,

I am using hash partitioning for both before and after dataset , I used sort stage , and same partitioning keys for sort and change capture , the meta data is same for both , though the data has not been changed but I am getting updates ,

pl suggest what may be the cause ,

thanks

[kumar_s]

Hi,

Try to find

Code: Select all

diff file1 file1

from unix.
It will reveal the uniquness of the files.

-Kumar