mixted data types for the same variable

[jdk123]

dataset1:

2001 $500 $400 200 300
2002 $600 $300 100 200


After transpose:

dataset2:

_name_ 2001 2002

acct1 $500 $600
acct2 $400 $300
acct3 200 100
acct4 300 200


as you can see from dataset2 two different data types asscociated to the year variables.

I would like to perform some calculations for instance: acct1 $600-$500=$100;
acct3 100-200 =100. is this possible, if yes, how to handle this situation?

Thanks,

[Cynthia_sas]

Hi:
If you do a PROC CONTENTS on dataset2:
[pre]
proc contents data=dataset2;
run;
[/pre]

and then look at the list of variables, what is the type and format for the "year" variables???? SAS variable names cannot start with a number, they generally start with an underscore from transpose, such as _2001 and _2002. Can you paste the variable information to the forum???

cynthia

[jdk123]

the variable name is _2001 _2002

[Cynthia_sas]

and the PROC CONTENTS???

[Ksharp]

It is possible. $w.d is only the cloth of variable ,in sas it is numeric.
if you want different cloth for the same variable ,you can create another variable to contain the format then use function putn() or putc() to get it.

[pre]
_name_ 2001 2002 formate

acct1 $500 $600 dollar8.
acct2 $400 $300 dollar8.
acct3 200 100 8.
acct4 300 200 8.


[/pre]


Ksharp