SAS: How do i split by comma and transpose? - sas

Question

I am using SAS Enterprise Guide.
I have a new file and i was asked to generate output.
Source:
Name feeder_in feeder_out NickName
ABBA 1,2 A,B ABBA
POLA 1,2 C,D,E CONS POLA
and the desire output:
Name feeder_final
ABBA 1
ABBA 2
ABBA A
ABBA B
POLA 1
POLA 2
CONS POLA C
CONS POLA D
CONS POLA E
I have been trying myself on handling this but no luck at all.
I tried
data test;
catequipment=catx(',',strip(feeder_in),strip(feeder_out));
do i=1 to countw(catequipment,',');
catequipment=catx(',',strip(feeder_in),strip(feeder_out));
do i=1 to countw(catequipment,',');
output;
end;
xequipment=newequipment;
run;
Does anyone have clue for this?

Answer 1

Here's my understanding of your requirements, based on the desired output: you want your output to have one observation for each combination of NAME and FEEDER_IN, plus another observation for each combination of NICKNAME and FEEDER_OUT.
On that assumption, the code would look something like (not tested):
data want;
set have;
keep name feeder_final
* Loop over FEEDER_IN and output one obs for each delimited value;
do i = 1 to countw(feeder_in, ',');
feeder_final = scan(feeder_in, i, ',');
output;
end;
* Move the NICKNAME value into NAME;
name = nickname;
* Loop over FEEDER_OUT and output one obs for each delimited value;
do i = 1 to countw(feeder_out, ',');
feeder_final = scan(feeder_out, i, ',');
output;
end;
run;

Answer 2

When transposing multiple columns you might want to also maintain the source row and column identifiers for further downstream analytics. The sequence of the values in the csv might also be important if you need to do pairwise joining on sequence position of the categorical form -- such as needing to match 1A 2B in row 1 and 1C 2D in row 2.
data have;
length name feeder_in feeder_out nickname $20;
input
Name& feeder_in& feeder_out& NickName&; datalines;
ABBA 1,2 A,B ABBA
POLA 1,2 C,D,E CONS POLA
run;
data want;
_row_ + 1;
set have;
feeder = 'in ';
do seq = 1 to countw(feeder_in,',');
value = scan(feeder_in,seq,',');
OUTPUT;
end;
feeder = 'out';
do seq = 1 to countw(feeder_out,',');
value = scan(feeder_out,seq,',');
OUTPUT;
end;
keep _row_ Name feeder seq value NickName;
run;