Hi Michele,

Sorry I am new to the SAS-L list. I tried to e-mail a response to the list but it did not get posted. I found today that you have to respond through the Internet "Compose a new message" screen. Here is what I had tried to send on 3/8/01.

When I first read your message I thought you needed to rotate an entire SAS table and make the variables observations and the observations variables. I was going to suggest Proc Transpose. After I looked at the results you were after I noticed that this looked familiar.

I did something similar to this recently. I wanted to put the results into a Proc Tabulate. Here is some code that might help. Basically you use an OUT= in your freq and then assign the counts to a variable. I used a macro and called it multiple times because each table of the freq would overwrite the previous one unless you output each table to another table. Then I appended the temp table to a master table and ran a Proc Tabulate on that. I tried this with one of your variables and it worked. You will just have to add macro calls for each of the variables that you want to add.

data test ; length gender $1 retail $1 ; input gender retail ; cards ; M C M C P C M R M R M R Z R P R P R P R P R ; run ;

proc freq data=test ; tables gender*retail / missing norow nocol nopercent ; run ;

proc datasets library=work ; delete mtemp ; quit ; run ;

%MACRO frqcount (var) ;

Proc freq data=test noprint ; tables &var*retail /list nopercent missing out=temp ; run ;

data temp ; length var $8 ; set temp (drop=percent rename=(&var=varval)) ; varname = "&var" ; run ;

proc append base=mtemp data=temp ; run ;

%MEND frqcount ;

%FRQCOUNT (gender)

proc tabulate data=mtemp format=8.0 missing ; class varname retail varval ; var count ;

table varname='VARIABLE' all ='TOTAL FOR ALL VARIABLES' ,

retail*varval='FLAG' * count=' ' * sum=' ' all ='ALL VALUES' * count=' ' * sum=' ' ; run ;

----------------------------------------------------------------------------

-----Original Message----- From: Michele Scherneck [mailto:mscherne@ROCHESTER.RR.COM] Sent: Wednesday, March 07, 2001 10:10 PM To: SAS-L@LISTSERV.UGA.EDU Subject: Report - rotated table

Hi everyone.

I've been working with reports and I'm having trouble developing a report that looks like the following examples. In the examples, retail_ flag is categorical. Gender, State and Income contain formatted values for data checking (they're a mixture of character and numeric). I could write a PROC FREQ, like this:

proc freq formchar (1,2,7)=' '; tables _all_*retail_flag/missing nocol norow; format _character_ $pop. _numeric_ pop.; format retail_flag; run;

The output would be similar to Example 1, but what I really want is something that looks like Example 2. Any ideas about how to rotate the data into that table format?

Example 1 Retail Flag Variable | C R --------------------|-------------------------------------- Gender Missing | 2 9 Zero | 0 0 Populated | 6 3 State Missing | 4 7 Zero | 0 0 Populated | 5 2 Income Missing | 2 6 Zero | 6 3 Populated | 4 1

Example 2

Retail Flag ------------------------------------------------------------------ | C | R --------------|-------------------------|------------------------- Variable | Missing Zero Populated | Missing Zero Populated --------------|-------------------------|------------------------- Gender | 2 0 6 | 9 0 3 State | 4 0 5 | 7 0 2 Income | 2 6 4 | 6 3 1

Many thanks, michele