topic List duplicate observations? in SAS Programming

List duplicate observations?

ybz12003 — Mon, 10 Jul 2017 15:55:47 GMT

Hello:

I would like to list the duplicate observations in 'Found' column. Please advice how. Thanks.

Best,

data Founddup;

informat name $80.;

input name $ found;

cards;

If_True 1

If_True_kary 1

If_True_John 3

If_Not 24

If_Not_Carol 24

If_False_Joe 288

;

Re: List duplicate observations?

art297 — Mon, 10 Jul 2017 16:07:24 GMT

data want;
  set Founddup;
  by found notsorted;
  if not(first.found and last.found);
run;
proc print data=want;
run;

Art, CEO, AnalystFinder.com

Re: List duplicate observations?

Reeza — Mon, 10 Jul 2017 16:09:38 GMT

NOUNIQUEREC option in PROC SORT does exactly this. You can also use the NOUNIQUEKEY if you're looking at specific variables to identify duplicates.

/*This code demonstrates how to keep only duplicate observations in a data set*/

%*Create sample data set;
data have;
informat name $80.;
input name $ found;
cards;
If_True 1
If_True_kary 1
If_True_kary 1
If_True_John 3
If_Not 24
If_Not 24
If_Not_Carol 24
If_Not_Carol 24
If_Not_Carol 24
If_False_Joe 288
If_False_Joe 288
;
run;

%*Sort with NOUNIQUEREC option;
proc sort data=have out=duplicates  nouniquerec;
by name found;
run;

https://gist.githubusercontent.com/statgeek/1ff5a4aaca9b3f875d4defdaa598ae5e/raw/87a7485a46add02b5f0f0a96f21933a3f2ebd8be/proc_sort_output_duplicates.sas

Re: List duplicate observations?

ybz12003 — Mon, 10 Jul 2017 17:19:45 GMT

Just curious, is there a way that sort the 'found' duplicate obs after the 'not (first and last)' statement? Thanks.

Re: List duplicate observations?

art297 — Mon, 10 Jul 2017 17:37:06 GMT

Not sure what you're asking. My suggestion was based on the existing order of the records. If you needed it sorted, I'd go with the proc sort nouniquerec option that @Reeza suggested.

Art, CEO, AnalystFinder.com

Re: List duplicate observations?

ybz12003 — Mon, 10 Jul 2017 18:04:16 GMT

Well, the proc sort code doesn't work in my actual dataset.. Yours works. Basely, I would like to sort the dupicate numbers from zero to largest. I found your code doesn't come with this function. I need to add proc sort on more step. I wish I could do it in one data steps.

data want;

set Founddup;

by found notsorted;

if not(first.found and last.found);

run;

proc sort data=want; by found; run;

Re: List duplicate observations?

Reeza — Mon, 10 Jul 2017 18:56:09 GMT

@ybz12003 wrote:

Well, the proc sort code doesn't work in my actual dataset..

It's not clear who you're responding to, please quote the original post in your response.

PROC SORT will work for your situation in a single step. If it doesn't you're doing something wrong.

Re: List duplicate observations?

art297 — Mon, 10 Jul 2017 19:04:02 GMT

Couldn't you just use?:

proc sort data=have out=duplicates  nouniquerec;
  by found name;
run;

Art, CEO, AnalystFinder.com