PROC Transpose by YearMonth and ID for duplicate values - sas

Question

I'm trying to transpose this table:
YearMonth ID Purchase Purchase Value
201912 1 Laptop 1000
202012 1 Computer 2000
202112 1 Phone 1000
201912 2 Stereo 500
202012 2 Headset 200
To look like this using PROC Transpose:
ID Purchase_201912 Purchase_202012 Purchase_202112 PV_201912 PV_202012 PV_202112
1 Laptop Computer Phone [phone]
2 Stereo Headset - 500 200 -
I think I'll have to transpose multiple times to achieve this. The first transpose I've tried doing is this:
proc transpose data=query_final out=transpose_1 let;
by yearmonth agent_number;
run;
but I keep getting the error
ERROR: Data set WORK.QUERY_FINAL is not sorted in ascending sequence. The current BY group has YearMonth = 202112
and the next BY group has YearMonth = 201912.
I've checked the the data from the table I'm pulling from is indeed sorted in ascending order by YearMonth then grouped by agent number, so I'm not sure what this error is referring to. Could it be that not all IDs have the same YearMonths associated with them (i.e. in example above, ID 2 did not purchase anything in 2021).

Answer 1

Assuming agent_number is equivalent to id in your example, I reproduced the data:
data have;
infile datalines4 delimiter="|";
input yearmonth id purchase :$8. PV;
datalines4;
201912|1|Laptop|1000
202012|1|Computer|2000
202112|1|Phone|1000
201912|2|Stereo|500
202012|2|Headset|200
;;;;
quit;
You can use two proc transpose and then merge the latter
proc transpose data=have out=stage1(drop=_name_) prefix=Purchase_;
by id;
var purchase;
id yearmonth;
run;
proc transpose data=have out=stage2(drop=_name_) prefix=PV_;
by id;
var PV;
id yearmonth;
run;
data want;
merge stage1 stage2;
run;
Resulting in the desired output
id purchase_201912 purchase_202012 purchase_202112 PV_201912 PV_202012 PV_202112
1 Laptop Computer Phone [phone]
2 Stereo Headset 500 200 .
PS: In order to avoid getting the error you report, sort the data first, in the same manner as in the by statement in the proc transpose. However, it is not needed here as it is already sorted by id.