topic Re: Keep first and last row in SAS Programming

Keep first and last row

zz — Fri, 23 Sep 2016 14:47:45 GMT

Hello,

I have the following row of data:

Sub Date Phase

01 2016-06-06 Before

01 2016-06-24 During

01 2016-07-01 During

01 2016-08-19 During

01 2016-08-26 After

I need to keep the first and last During to have the data look like this:

Sub Date Phase

01 2016-06-06 Before

01 2016-06-24 During

01 2016-08-19 During

01 2016-08-26 After

I tried to use the first. and last variable like below:

DATA JR.SV_021;
SET SV_02;
BY PHASE;
IF first.PHASE THEN OUTPUT;
IF last.PHASE THEN OUTPUT;
run;

but it does not work for me. I appreciate your help very much in advance!

Re: Keep first and last row

Astounding — Fri, 23 Sep 2016 14:52:37 GMT

A small change should get this to work. Change the BY statement to:

BY PHASE NOTSORTED;

That will permit your BY statement even though the data are not in order by Phase.

Also note, if your actual data set is larger and might contain more than one SUB value, you may need to use:

BY SUB PHASE NOTSORTED;

Either way, the rest of the program can remain the same.

Re: Keep first and last row

zz — Fri, 23 Sep 2016 15:14:28 GMT

Tried, but somehow, it did not work:(

Re: Keep first and last row

ballardw — Fri, 23 Sep 2016 15:18:13 GMT

Describe how "doesn't work": Error, unexpected output, no output.

And show the code ran and the log.

Re: Keep first and last row

zz — Fri, 23 Sep 2016 15:35:44 GMT

Hi ballardw, It worked! I checked and one libref was not correct for input data. Once that's corrected, it all worked. Thank you very much!

Re: Keep first and last row

zz — Fri, 23 Sep 2016 16:20:52 GMT

Hi ballardw,

One problem, rows are duplicated, if phase changed in the next row, eg: the first row was duplicated.

Re: Keep first and last row

ballardw — Fri, 23 Sep 2016 16:44:39 GMT

Replace

IF last.PHASE THEN OUTPUT;

with

IF last.PHASE and not (First.Phase) THEN OUTPUT

Re: Keep first and last row

Saidurga1 — Wed, 26 Jul 2017 09:18:32 GMT

data durga;

input Sub 2. Date:yymmdd10. Phase $ 24-30;

infile datalines;

cards;

01 2016-06-06 Before

01 2016-06-24 During

01 2016-07-01 During

01 2016-08-19 During

01 2016-08-26 After

;

run;

proc sort data=durga out=class;

by Phase;

run;

data durga_;

set class;

by phase;

if first.Phase=1 or last.phase=1;

run;