topic Deleting Certain Categories within a Variable in SAS Programming

Deleting Certain Categories within a Variable

MN — Sat, 23 Jun 2012 06:26:00 GMT

Hello,

I have a categorical variable, labelled 0-9. I need to delete certain values so I can run a CHISQ.

This is what I'm trying, but with no luck:

DATA mn.lahrt; *my dataset*

SET IHD_DX; *the variable name*

IF IHD_DX = 0 THEN delete;

IF IHD_DX = 8 and 9 Then delete;

RUN;

I need to keep categories 1-7 only (which are further split in only two categories).

Any suggestions?

Re: Deleting Certain Categories within a Variable

shivas — Sat, 23 Jun 2012 06:58:37 GMT

Hi ,

Try this..Hope it helps

DATA mn.lahrt;

SET mn.lahrt;

IF IHD_DX in ('0','8','9') THEN delete;

RUN;

Thanks,

Shiva

Re: Deleting Certain Categories within a Variable

Tom — Sat, 23 Jun 2012 18:27:17 GMT

Your language is very confused. Sounds to me like you have a variable named IHD_DX in a dataset named MN.LAHRT.

You would not normally want to overwrite your input dataset as in your example code. In fact to limit data you can just use a WHERE statement in your procedure calls.

proc freq data=mn.lahrt ;

where 1<= idh_dx <= 7 ;

tables idh_dx / chisq ;

run;

I am not sure what you mean by "labeled" values. Is it possible that the actual values in IDH_DX variable are something other than the values 0,1,....9 ? Do you have a format attached?

You can also use formats to temporarily regroup you data for a PROC FREQ call. For example if you want to regroup your values into two groups you might use something like:

proc format ;

value group 1-7 = 'VALID' other='INVALID';

run;

proc freq data=mn.lahrt;

format idh_dx group.;

tables idh_dx;

run;

Re: Deleting Certain Categories within a Variable

MN — Sun, 24 Jun 2012 07:34:42 GMT

Thank You for your help. Though that didn't really work for my data-set, so I just exported it to excel and made the needed changes.

Re: Deleting Certain Categories within a Variable

MN — Sun, 24 Jun 2012 07:35:03 GMT

Thank You for your help. Though that didn't really work for my data-set, so I just exported it to excel and made the needed changes.