topic Re: Obtain the latest collection date in Proc Freq or from extra Data steps? in SAS Programming

Obtain the latest collection date in Proc Freq or from extra Data steps?

ybz12003 — Thu, 18 Jul 2024 19:17:24 GMT

I have a PROC FREQ code listed below in the frequency output dataset, CNT. I would like to obtain information for the latest collection date in the "Have" dataset. Is there a way to add a step in the PROC FREQ procedure, or do I need to add an extra data step and merge it with the output CNT data? Please help, thanks.

Proc freq data=Have noprint; 
	table  LabIDs*Test_Code*Match / nopercent nocol  out=CNT;
	where match in ('N');
run;

Re: Obtain the latest collection date in Proc Freq or from extra Data steps?

mkeintz — Thu, 18 Jul 2024 19:24:59 GMT

You do not provide enough information to answer your own question.

Mainly, what exactly do you mean by "collection date"? Is it a variable in the dataset HAVE? You specify three variables in HAVE (Labids, test_code, match), none of which suggest a collection date.

Is it the creation date of HAVE?

BTW, there is no need for the ** match" expression in the TABLE statement, since you allow only one match value, i.e. where match in ('N').

Please clarify.

Re: Obtain the latest collection date in Proc Freq or from extra Data steps?

ballardw — Thu, 18 Jul 2024 19:53:50 GMT

@ybz12003 wrote:

I have a PROC FREQ code listed below in the frequency output dataset, CNT. I would like to obtain information for the latest collection date in the "Have" dataset. Is there a way to add a step in the PROC FREQ procedure, or do I need to add an extra data step and merge it with the output CNT data? Please help, thanks.

You ask "for the latest collection date in the "Have" dataset" but your code is getting counts by LabIds and Test_code. So should that request be "for the latest collection date in the "Have" for each LabIds and Test_code combination in the dataset"? Quite a bit different things.

If you want latest date for each combination of LabIds and Test_code AND have a collection date variable one way AND that the sort order for the collection date variable sorts in date order (i.e and an actual SAS date value or a YYYYMMDD value) the following should work.

Proc freq data=Have noprint; 
	table  LabIDs*Test_Code*collectiondate*Match / nopercent nocol  out=CNT;
	where match in ('N');
run;

data want;
   set count;
   by Labids test_code;
   if last.test_code;l
run;

Re: Obtain the latest collection date in Proc Freq or from extra Data steps?

ybz12003 — Thu, 18 Jul 2024 19:58:33 GMT

The Collection_Date volume (mmddyy10.) is in the "Have" dataset, but not in the CNT frequency output dataset.