CROSS DATABASE KEY ISSUE

[HeatherNewton]

I am loading sas data converted to csv into DB2. one problem is when  I setup tables in DB2, I added keys and the keys are usually the merge by columns. and these keys are not nullable. So if in sas data, key is null then when I load into DB2, the entries with null key would be rejected automatically. And my uat and prod data might not match. how can I overcome this? is it a good way if I transferred null to empty string so it wouldnt be rejected by db2?

[Kurt_Bremser]

"Empty" is the same as "missing" for SAS character variables.

Today, most keys are UUIDs, and you can create such with the UUIDGEN function; format them to the readable 36-character notation for the csv file.

[LinusH]

If we are talking about strings, yes it could be done this way.

Remember that the key still needs to be unique.

I think yuo should take care of your data quality problem first before loading to DB2.