Difference between NODUPKEY and NODUPRECS?

[赵家钧]

I'm wondering if there exists any difference between NODUPKEY and  NODUPRECS options ? In my view, they seem totally the same with a BY statement.

[VDD]

NODUP will delete duplicated observations (records that are identical to each other) while NODUPKEY will delete those observations that have duplicate BY values (the sort BY variables that you name in PROC SORT).

Links to SAS papers can be found here
https://communities.sas.com/t5/SAS-Programming/Difference-between-NOdup-and-NoDupkey/m-p/29490#M6614...

[赵家钧]

But..... NODUPRECS option also has a BY statement. Aren't they the same?

[kiranv_ · Accepted solution]

go through this article, it explains everything step by step with examples.

https://support.sas.com/resources/papers/proceedings/proceedings/sugi30/037-30.pdf

[赵家钧]

I think I get the answer just after reading the ABSTRACT!!! Thank you a lot! I just wrongly thought that the BY statement of NODUP can be customized depending on my needs.

[Reeza]

NODUPRECS has been removed from PROC SORT (documentation) but it still works for backwards compatibility. I would not recommend developing any programs that rely on the functionality of NODUPRECS today.

If you've read the paper, there are some gotcha's about it that often caught people, so they changed the functionality.

[agoldma]

In addition to Reeza, another person from SAS Support also said that NODUPRECS has been removed from the most recent documentation (9.4) because it works in an unexpected/unreliable way.

Instead of NODUPRECS, the current suggestion is to use NODUPKEY and to sort BY _ALL_