topic Re: Distinct row numbers data step in SAS Programming

Distinct row numbers data step

Q1983 — Mon, 07 Oct 2019 19:11:06 GMT

data test;

input ln $ step $ date1 date9.;

format date1 date9.;

datalines;

0122 B 01jun2019

0122 B 07jun2019

0122 C 06jun2019

0133 C 07jun2019

0133 B 01jun2019

;

run;

The desired output is to get distinct row numbers based on the change in ln,step and date. Then reset with a new step or ln. Something like this;

Shmuel — Mon, 07 Oct 2019 20:08:14 GMT

You asked for "distinct row numbers based on the change in ln,step and date." but the result shows distinct row per ln step

only, in 0122 B 07jun2019 - why is row_num = 2 ?)

Any way, assuming data is sorted by ln step then:

data want;
 set have;
     by ln step;
        retain row_num;
        if first.step then row_num=1; 
        else row_num+1;
run;

ChrisNZ — Mon, 07 Oct 2019 20:06:27 GMT

Your explanations are unclear to me.

This matches your output:

data WANT; 
  set HAVE;
  if lag(DATE1)>=DATE1 then ROW_NUM=0;
  ROW_NUM+1;
run;

LN	STEP	DATE1	ROW_NUM
0122	B	01JUN2019	1
0122	B	07JUN2019	2
0122	C	06JUN2019	1
0133	C	07JUN2019	2
0133	B	01JUN2019	1
0133	B	01JUN2019	1

PGStats — Tue, 08 Oct 2019 03:38:07 GMT

It looks like ln has no role in determining row_num, so I commented it out:

proc rank data=test out=want ties=dense;
by /*ln*/ step notsorted;
var date1;
ranks row_num;
run;