Change capture not taking input from "before" link

[ady]

The 2 sources I am using are sequential files with diferrent names, they have the same data and metadata.

I am just trying to test the stage as I'v never used it before.

There is 1 key for the file and I am doing an Internal ascending sort.

change mode = Explicit keys, All values

Drop Output for Copy = False

Drop Output For Delete= False

and everything in the options is marked False....

when I am running the job I see that rows are comin in from just the "after" link, No rows are coming form the "before" link.

What am I doin wrong here ?

[us1aslam1us]

Have you defined both the files in the Input tab of Change Capture stage?

sam

[chulett]

So why is this marked as a 'Server' question?

[ady]

Do you mean the METADATA ? .. I havnt given the name of the files anywhere in the CC stage. Am i required to ?

The link ordering tab shows which file is input and output.

[ady]

So why is this marked as a 'Server' question?

oops my bad ..

[ady]

Error: Input 'cc_PromoMaster.l_Previous' Sort requires a key column
Error: Input 'cc_PromoMaster.l_Current' Partitioner requires a key column
Error: Input 'cc_PromoMaster.l_Current' Sort requires a key column


Whats does this mean ?... I have specified a key

[ray.wurlod]

Go to the Inputs tab, to the Partitioning tab thereon. Have you specified a sort key there? For BOTH inputs? (Make sure, too, that you have specified identical key-based partitioning for both inputs.)

[ady]

Done, Works now ... 1 more question

As of now I am using "change code" to check if the data has been changed. I wanna know how to write the changed data to a file.

I have created test data to check if what I am doin is right, I made changes to 3 columns in one of the source data by hardcoding a value.
But when I make "Drop output for copy" = "true" and the rest of the options to "false" and run the job it gives out the it should only write out the 3 columns right??

But it writes all the data out !

[kumar_s]

The condition don't restrict any column, but only rows. You can tell us what are the input columns and what are the output.
You can check the 'Change code' for which the output is obtained.
You should get the output of after records for other than copy.

[ady]

I have about 60 columns , MGMT_NO is the key.

I have manually hardcoded the city name and zipext . So now when I select the option "Drop Output for Copy" to be "True" the job should write out the rows which have been hardcoded by me right ?

[kumar_s]

If you look in to the help for Change code, you can understand that each code have different meaning. Copy: both before and after key and values are same.
Delete: If Before key is not present in after key.
Insert : If its a new key came in After records.....

[ady]

I think I'v confused you guys.


The job I have to do is compare 2 sequential files (1 from DEV and 1 from PROD) and see if they are the same. If they are not the same and the records have different data then I have to write the data which is different into another Seq file.

I am trying to use Change capture for this, I need to know if I am missing something here

[kumar_s]

Ok going back to your question, if both the key fields are same and all the values specified are same, it comes through Copy change code, which includes city name and zipext.
If you are just testing it, allow all the type of change code and analyze what record came with which change code. It will give you an idea of what you expected and what went wrong (if any).

[ady]

I tried it and when I select TRUE for the DROP OUTPUT FOR COPY option , It gives me the correct result with the CODES.

But I want the job to output the rows which are not present in the BEFORE data set and are present in the AFTER data set. ie: the updated rows.

Is there a option to do that ?

[kumar_s]

As I mentioned earliar, the records that are not present in Before is INSERT. (NEW records).