topic Re: How to split DX for an encounter into multiple rows? in SAS Health and Life Sciences

How to split DX for an encounter into multiple rows?

renjithr — Mon, 21 Oct 2013 22:23:21 GMT

Hi All,

Could you please help me with this problem.

My file has the following fields:

ID rgn_cd dx1 dx2 dx3....dxn

100 01 200 300 400 ... 110

I want to create an output file like this:

ID rgn_cd dx1 dx2 dx3 ..... dxn

100 01 200

100 01 300

100 01 400

... --- ------

100(n) 01(n) 110

Please share your thoughts

Re: How to split DX for an encounter into multiple rows?

Reeza — Mon, 21 Oct 2013 22:28:07 GMT

Don't. That file structure makes very little sense, what would be the benefit?

A long structure is better, and you can use proc transpose to get it.

ID rgn_cd dx dx_code

100 01 1 200

100 01 2 300

...

100 01 n 110

101 01 1 201

Re: How to split DX for an encounter into multiple rows?

renjithr — Mon, 21 Oct 2013 22:32:31 GMT

The request was to modify the output file like I mentioned above to tally the dx for the client, I agree that long structure is better but that is not what my client wants.

Re: How to split DX for an encounter into multiple rows?

Reeza — Mon, 21 Oct 2013 22:33:17 GMT

How would you tally the DX for the client with that structure, compared to the long structure?

Re: How to split DX for an encounter into multiple rows?

renjithr — Mon, 21 Oct 2013 22:38:18 GMT

I think what the client wants is to count the rows for an encounter and compare that count with the total DX codes.

for eg: an ecnounter 101 has 4 dx codes then the output file will have 4 rows for that encounter.

Re: How to split DX for an encounter into multiple rows?

Tom — Mon, 21 Oct 2013 23:01:17 GMT

Use arrays and a DO loop.

%let n=10 ;

data want ;

set have;

array _dx dx1-dx&n ;

array save(&n) $5 _temporary_ ;

do _n_=1 to &n ;

save(_n_)=_dx(_n_) ;

end;

call missing(of dx1-dx&n);

do _n_=1 to &n ;

_dx(_n_)=save(_n_);

output;

end;

run;

Re: How to split DX for an encounter into multiple rows?

renjithr — Mon, 21 Oct 2013 23:27:44 GMT

Hi Tom,

Thanks so much for your input. With a slight modification I was able to get what I wanted.

Below is the modified code:

%let n=10 ;

data want ;

set have;

array _dx dx1-dx&n ;

array save(&n) $5 _temporary_ ;

do _n_=1 to &n ;

save(_n_)=_dx(_n_) ;

end;

do _n_=1 to &n ;

call missing(of dx1-dx&n);

_dx(_n_)=save(_n_);

output;

end;

run;