topic count function in data step in SAS Programming

count function in data step

Ross1010 — Wed, 17 Apr 2019 15:22:57 GMT

i'm new to sas so apologies if this is really easy but i want to total up the fields which have a condition in and those that have a blank in the field. I have 4 conditions so a way to work out the total for condition 1 with/without something in field, then condition 2, condition 3 etc

Hope that makes sense

Cheers

Code i have so far is

rsubmit &UNIX;
data pol_trv_dims;
set DLItm.pol_trv_dims
(keep = POLKEY CONDITION1 CONDITION2 CONDITION3 CONDITION4);

run;
endrsubmit;

Re: count function in data step

ballardw — Wed, 17 Apr 2019 15:51:49 GMT

@Ross1010 wrote:

hi

i'm new to sas so apologies if this is really easy but i want to total up the fields which have a condition in and those that have a blank in the field. I have 4 conditions so a way to work out the total for condition 1 with/without something in field, then condition 2, condition 3 etc

Hope that makes sense

Cheers

Code i have so far is

rsubmit &UNIX;
data pol_trv_dims;
set DLItm.pol_trv_dims
(keep = POLKEY CONDITION1 CONDITION2 CONDITION3 CONDITION4);

run;
endrsubmit;

It really helps to show some actual example input data and what the end result should be. Provide some records that will exercise each of your rules.

You may want to try:

proc summary data=pol_trv_dims;
   class condition1 - condition4 /missing;
   output out=work.summary;
run;

The output data set will have combinations of the variables on the class statement, including one record for "missing", and a count of records in the variable _freq_. There will be another variable _type_ that indicates which set of combinations each record represents,

Type=0 will a total count of records, type=1 will be the summary for one variable only, expect the last one on the CLASS statement, 2, 3 and 4 would be each of the other variables on the class statement alone, higher numbers will represent different combinations. The largest value of _type_ will be all of the variables combined.

You can filter the data for other use on the value of _type_ to get the records you are interested.

Re: count function in data step

Astounding — Wed, 17 Apr 2019 17:51:21 GMT

This might be a little more than you asked for, but it should all be useful:

proc format;
value $missfmt ' '='Blank' other='There';
run;

proc freq data=have;
tables condition1-condition4 / missing;
tables condition1*condition2*condition3*condition4 / list missing;
format condition1-condition4 $missfmt.;
run;