topic Deleting / Keeping a subset of observations in SAS Data Management

Deleting / Keeping a subset of observations

r4321 — Thu, 16 May 2019 02:41:36 GMT

Hello all,

Forgive me if my code is a bit elementary, but I code use some help with a simple data step question.

I am trying to keep a subset of observations in my data. I've often done this in the past by using if var = "..." then delete in a data step. However, if I did it that way, I would have far too many to delete. Alternatively, if i did it by using if var ne "..." then delete, I would have to use a bunch of data steps and then merge into a large file, which is also not super convenient. I'm wondering can I do something like

data want;
set have;
keep if var1 = "CharacterName";
keep if var1 = "CharacterName1";
keep if var1 = "CharacterName2";
keep if var1 = "CharacterName3";
run;

However, when I run this, it doesn't work of course.

Any suggestions?

Thanks in advance!

Re: Deleting / Keeping a subset of observations

Patrick — Thu, 16 May 2019 03:03:39 GMT

@r4321

The KEEP statement defines variables (columns) you want to keep in your output data set (the one defined using DATA ....). SAS tables are rectangular so it's either keep a column or don't. Also for this reason a KEEP statement is never conditional.

You want to remove rows. With SAS you normally simply create a new table (WANT in your case) and then select the rows you want.

Syntax you could use:

if var1 in ('CharacterName', 'CharacterName1',....);

Re: Deleting / Keeping a subset of observations

r4321 — Fri, 17 May 2019 20:21:12 GMT

Thank you!