Are there ANY adverse effects to using by statement's NOTSORTED option?

[art297]

I can't recall if this question has ever being asked. When one includes a by statement in a procedure, and follows the by variable names with the option NOTSORTED, and the data are sorted, the option doesn't appear to have any adverse effect.

Does it have any effect(s) I'm not aware of?

[PGStats]

Good question!

The option NOTSORTED cannot be used when the ordering of two or more datasets needs to be the same. You cannot use it in a datastep with MERGE and UPDATE. You can make NOTSORTED the default with global option NOBYSORTED.

More generally, I find the default option (that the observations are sorted by the BY variables) a useful safeguard against accidental reordering, as can happen for example as the result of a SQL query.

PG