topic Re: format in SAS Programming

format

Dhana18 — Fri, 06 Jul 2018 13:11:50 GMT

Data test;
Input Event_id Drug_use $10. Condom_use $10.;

Datalines;
123 Marijuana Always
456 Meth Sometimes
789 . Always

156 Heroin .
147 . .
269 Meth Never
run;

proc print data=test;

run;

I have this:

Event ID	Drug_use	Condom_Use
123	Marijuana	Always
456	Meth	Sometimes
789	.	Always
156	Heroin	.
147	.	.
269	Meth	Never

I want this:

Event ID	Drug_use	Condom_Use
123	1	1
456	1	1
789	0	1
156	1	0
147	0	0
269	1	1

Your guidance would be helpful.

Thank you

Re: format

Kurt_Bremser — Fri, 06 Jul 2018 13:20:44 GMT

Missing values in character variables are blanks.

Also make correct use of formats in list input.

Data test;
infile cards dlm=' ' dsd;
Input Event_id _Drug_use :$10. _Condom_use :$10.;
Drug_use = _Drug_use > '';
Condom_use = _Condom_use > '';
drop _Drug_use _Condom_use;
Datalines;
123 Marijuana Always
456 Meth Sometimes
789  Always
156 Heroin 
147  
269 Meth Never
run;

proc print data=test;
run;

Re: format

soham_sas — Fri, 06 Jul 2018 13:29:21 GMT

@Kurt_Bremser kindly please explain how drug_use and condom_use are assigned to 0 and 1

i didn't get it so asking your help

Re: format

Tom — Fri, 06 Jul 2018 14:00:20 GMT

Use the MISSING() function.

So given this test dataset

data test;
  input Event_id Drug_use :$10. Condom_use :$10.;
datalines;
123  Marijuana  Always
456  Meth  Sometimes
789  .  Always
156  Heroin  .
147  .  .
269  Meth  Never
;

If you just do this:

data want ;
  set test ;
  drug_use = not missing(drug_use);
  condom_use = not missing(comdom_use);
run;

You will have some trouble since those variables are already defined as character variables instead of numeric variables.

You can instead make new variables. If you want to use the same names then you will need use rename.

Here is one way.

data want ;
  set test ;
  drug_use_flag = not missing(drug_use);
  condom_use_flag = not missing(comdom_use);
  drop drug_use condom_use;
  rename drug_use_flag = drug_use condom_use_flag=condom_use ;
run;

Re: format

Dhana18 — Fri, 06 Jul 2018 14:39:15 GMT

Sorry for not being able to explain well for what I need. Ok let me try again.
I imported an excel file into SAS and it looks like this
Event_id

Risks

123
456
123
974
456

drug use Marijuana
drug use Heroin
condom use(sometimes)
drug use meth
condom use(always)

I transposed the data set now it looks like this
Event_id

drug use marijuana

drug use heroin

drug use meth

condom use(sometimes)

condom use(always)

123

drug use marijuana

condom use(sometimes)

456

drug use heroin

condom use(always)

974

drug use meth

So insteatd of this I want
Event_id

drug use marijuana

drug use heroin

drug use meth

condom use(sometimes)

condom use(always)

123

Yes

No

No

Yes

No

456

No

Yes

No

No

Yes

974

No

No

Yes

No

No

It can be yes or no or 1 and 0.
I really need your help please.

Re: format

Tom — Fri, 06 Jul 2018 14:49:28 GMT

Your latest example is unreadable. Please use the {i} icon to insert text so that spaces used for alignment are preserved.

Re: format

Dhana18 — Fri, 06 Jul 2018 15:05:54 GMT

I imported an excel file into sas, that looks like this

event_id risks

123 drug use marijuana

456 drug use heroin

123 condom use (sometime)

974 drug use meth

456 condom use(always)

Then I transposed the dataset now it looks like this in sas

event_id drug use marijuana drug use heroin condom use (sometimes) drug use meth condom use(always)

123 drug marijuana condom use (sometimes)

456 drug use heroin condom use(always)

974 drug use meth

I want the table like this

event_id drug use marijuana drug use heroin condom use (sometimes) drug use meth condom use(always)

123 1 0 1 0 0

456 0 1 0 0 1

974 0 0 0 1 0

Re: format

Tom — Fri, 06 Jul 2018 15:14:16 GMT

I am pretty sure it doesn't look like:

event_id risks
123  drug use marijuana
456 drug use heroin
123 condom use (sometime)
974 drug use meth
456 condom use(always)

Perhaps it looks like the dataset you would get if you ran this code:

data have ;
  length event_id 8 risks $30 ;
  infile cards dsd dlm='|' truncover ;
  input event_id risks;
cards;
123|drug use marijuana
456|drug use heroin
123|condom use (sometime)
974|drug use meth
456|condom use(always)
;

You cannot really make a dataset with variable names like in your wanted output. Perhaps you want a report instead?

To get a report like you describe why not just use PROC FREQ?

proc freq data=have ;
 tables event_id*risks ;
run;

Re: format

Reeza — Fri, 06 Jul 2018 15:18:03 GMT

It looks like what text is there doesn't matter, just that there's text. And then you convert any text to 1 and 0 otherwise.

Is that correct?

If so, declare two arrays, one that has all your original and one that has your new variables.

Then loop through and assign your 0/1.

Untested code below:

array old_vars(*) $50 use_marijuana use_heroin condom_sometimes condom_always;
array new_vars(*) drug_marijuana drug_heroin condom_1 condom_2;*variable names need to be different;

do i=1 to dim(old_vars);
if missing(old_vars(i)) then new_vars(i) = 0; else new_vars(i)=1;
end;

Re: format

Kurt_Bremser — Fri, 06 Jul 2018 16:56:57 GMT

@soham_sas wrote:

@Kurt_Bremser kindly please explain how drug_use and condom_use are assigned to 0 and 1

i didn't get it so asking your help

The result of a comparison is a boolean value, 0 for false and 1 for true.

x = y = z;

would mean: if y equals z, assign 1 to x, otherwise assign 0.