| Date: | Mon, 26 Oct 2009 12:52:19 -0400 |
|---|
| Reply-To: | Charles Harbour <charles.harbour@ACT.ORG> |
|---|
| Sender: | "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU> |
|---|
| From: | Charles Harbour <charles.harbour@ACT.ORG> |
|---|
| Subject: | Re: How can we stop the "do nothing data step"? |
|---|
If it can't be done in 80 columns, it's not worth doing. :-)
I suspect y'all are on the right path about the do-nothing datastep. In
structured languages (cobol, fortran), you typically have to load your data
into working storage. And, as has already been pointed out, SAS will
convert your sequential tape dataset to random access on dasd, which is <so>
much easier (can't tell you how many 'you're using the sequential engine'
errors I've received over the years)(Bad programmer, no data for you!)
20 years from now, they'll be puzzling over your code, too!
CH
|
