changing columns into rows

[Linlin]

Good morning Art!

I am using a Lenovo laptop with Windows 7.

[rtritz]

Hi Art and Linlin,  I have Sony laptop with Windows Vista.

Afterwards, I was wondering about something like this or a system option that may have a different value.

Rich

[art297]

Rich,

What do you get if you run:

proc options option= caps;

run;

I ran your code on 9.2 and got the same result as Linlin did.

[rtritz]

Art,

Proc options option=caps;

run;

gives NOCAPS in the log.

I was looking at VALIDVARNAME could that have any effect?

Rich

[rtritz]

Hi Art and Linlin,

The VALIDVARNAME option may be the key.

The setting for this on my system is UPCASE.

From the documentation validvarname=V7 is the default and UPCASE and V7 essentially do the same things as far

as having variable names be all uppercase in SAS datasets. I changed the value to validvarname=any(from 9.3 documentation) and two things happened.

I got a Warning in the log saying this setting was experimental in 9.1.3 but when I did this then the variable names in the dataset

I tested were the same case as written in the code and no longer all uppercase.

Could ANY now be the default for 9.2 and later or is it still V7? Could you check this setting on your systems?

Rich

[art297]

9.2: validvarname=v7

[Linlin]

Hi Rich,

from my log file:

2806  Proc options option=VALIDVARNAME;

2807  run;

    SAS (r) Proprietary Software Release 9.3  TS1M0

VALIDVARNAME=V7   Control the type of SAS variable names that can be created

                   during a SAS session

NOTE: PROCEDURE OPTIONS used (Total process time):

      real time           0.00 seconds

      cpu time            0.00 seconds

[rtritz]

Alright, that appears to explain it. When I change my VALIDVARNAME option to V7 then the variable names in the dataset

I tested were the same case as in the code and not uppercase anymore. So, when and if I ever post to the forum

again, I will have to think about changiing that option especially if I quote any variable strings.

Rich

[dan999]

This worked. Thanks a lot.

[MikeZdeb]

hi ... another idea ...

data x;

input program  col1-col3 @@;

datalines;

10 3  5  1 15 1  2  7  24 6  8  5 21 2  8  1 35 9  4  3 33 3  1  1

;

proc format;

value p

10,15 = 'A'

24,21 = 'B'

35,33 = 'C'

;

run;

proc summary data=x nway;

class program;

var col: ;

output out=stats sum=;

format program p.;

run;

proc transpose data=stats out=want (drop=_name_) prefix=prg;

var col1-col3;

id program;

run;

Obs    prgA    prgB    prgC

1       4       8      12

2       7      16       5

3       8       6       4