topic Re: Comparing counts of different values of a variable present in two datasets. in SAS Programming

Comparing counts of different values of a variable present in two datasets.

niharik — Fri, 23 Sep 2022 17:39:50 GMT

Hello Friends,

I have a variable X having values (.,Yes, No). I have this variable in two different datasets A and B with an ID. I have to find the frequency matrix of count of values in both datasets as below. Please help me how it is possible either with Proc freq or any other procedure.


X	datasetA
datasetB Frequency Percent Row Pct Col Pct	.	No	Yes	Total
.	6764 63.52 99.75 99.73	12 0.11 0.18 0.47	5 0.05 0.07 0.39	6781 63.68
No	14 0.13 0.54 0.21	2560 24.04 99.15 99.46	8 0.08 0.31 0.62	2582 24.25
Yes	4 0.04 0.31 0.06	2 0.02 0.16 0.08	1280 12.02 99.53 98.99	1286 12.08
Total	6782 63.69	2574 24.17	1293 12.14	10649 100.00

Re: Comparing counts of different values of a variable present in two datasets.

PaigeMiller — Fri, 23 Sep 2022 17:49:10 GMT

In order to get any analysis like this, you need the variables of interest in one data set; not separated in two data sets.

To combine the data sets, you would have to do a MERGE in a SAS data step, by ID, and also rename the variables so they both don't have the same name. Something like this

/* UNTESTED CODE */

data combined;
    merge a(rename=(x=x_a)) b(rename=(x=x_b));
    by id;
run;

For this to work, both data sets must be sorted by ID. Then PROC FREQ can take data set COMBINED and produce the report you want.

Re: Comparing counts of different values of a variable present in two datasets.

niharik — Fri, 23 Sep 2022 19:18:50 GMT

Thank you. It worked after renaming, merging and running proc freq between renamed vars.