Summarizing long file

[Walternate]

Hi all,

I have a long dataset that contains file "pairs." Some pairs have identical names, others have similar but not exactly the same.  I have the date, name, and ext for each file:

FNM1       FDT         FEXT          FNM2      FDT2       FEXT2

abc          1/1/19          csv            abc           1/1/20     XLS

abc          1/1/19          csv            abc_d       2/1/20    CSV

def          2/1/19          csv            def            3/1/20     XLS

ghi           3/1/19          csv            ghi            4/1/20     XLS

ghi           3/1/19          csv            ghi             5/1/20    XLSX

What I'm trying to do is flag weird cases for examination. A "normal" case that I would expect to see would be:

                 F1 is a CSV 

                  F2 is an XLS with the same name (dates don't really matter here)

If anything else is going on, I'd like to flag for examination, so for example:

               1) F1 and F2 are both CSVs (these would have slightly different names)

               2) F1 is CSV and F2 is XLSX (same name)

So basically what I'm looking to do is build an output so I can examine all cases that are not the "normal" case that I outlined above.

Any  help is much appreciated.

[ballardw]

And the output should look like what exactly?

[A_Kh · Accepted solution]

Based on provided condition, this code will flag it. 

data want;
	set have;
	length case $10;
	if strip(upcase(fnm1)) eq strip(upcase(fnm2)) then do;
		if upcase(strip(fext)) eq 'CSV' and upcase(strip(fext2)) eq 'XLS' then Case='Normal';
		else if upcase(strip(fext)) eq 'CSV' and upcase(strip(fext2)) eq 'XLSX' then Case='Not normal';
	end;
	else Case='Not normal';
proc print;run; 

[Quentin]

Sounds Like:

data want;
  set have;
  Flag = NOT (FEXT='csv' and FEXT2 ='XLS' and FNM1=FNM2) ;
run;

Flag will be 0 when a record meets the 'normal' condition, and will be 1 when a record is not normal.

This is like an assertion.  You write an expression you expect to be true, and flag the records where it is not true.