DSXchange

Hi Gurus

I am generating a CSV file from within a job with sequential file stage.However , whenever there is no data at the source side , it is generating a blank CSV file. Is there any way I can prevent this or any workaround .

Prevent? No. If it really bothers you, you can build an after-job subroutine to detect the empty file and delete it if you like.

kasgang wrote: I am generating a CSV file from within a job with sequential file stage.However , whenever there is no data at the source side , it is generating a blank CSV file.

Isn't it supposed to??

kasgang wrote: Is there any way I can prevent this or any workaround .

Why would you want to prevent this?

Hi,
By definition an empty fie should contain nothing if it is really empty
This is done as part of the overwrite operation on the previos file.

You can try to delete any existing file, then only run the job if there are rows to process.
Or you can delete the file after verifying no rows were written to it.
IHTH,

roy wrote:By definition an empty fie should contain nothing if it is really empty

Or it could contain a single header record. That's why, IMHO, you should check the LinkRowCount and not the file size as that will 'always' work for you.

OK I ment DATA wise Craig, probably that exclamation mark was uncalled for
Craig if you were refering to a file header record that contains some general and or validation data like the number of rows that the file should contain then that is also data.

Obviously the way to determine number of rows is subject to the process logic and should reflect a proper method to reflect existance of or lack of rows

Hi Gurus

Thanks for the inputs. By empty files I meant , file with header rows only . The way I am doing it now , is checking the LinkRowCount and writing to another file the filenames with LinkRowCount=0.
Then I am deleting the empty CSV files referring to this file.

Just wanted to know if there is anythig better that way to do the same.

I guess the only 'better' way would be to stop worrying about it and just leave them alone.

Make use of Folder stage as the target stage instead of the sequential file stage and your requirement will be achieved easily. we have done the same things successfully.