optimisation of the sas table before reporting

[Mic35]

Hello,

After doing a lot of research, I would like a quick help please....

I am looking to optimise my table for reporting purposes (please see attachment). 
The aim is to lighten the table of duplicate data. 
I tried different solutions with data step, without success.
My problem can be fixed?

The table is the result of a succession of macros and it's difficult to solve it before the output. 
Modifications are easy in excel on a variable but when there are many of them, it's nicer to automate and it reduces the risk of error.

I appreciate your help and tips.

Mic

[TomKari]

Hi!

How are you creating that result, and what format is it in?

Is there a SAS dataset that we could work with, or do your macros write directly to a report. If there's a SAS dataset, what does it look like?

Can you describe the process of creating it in a bit more detail?

I can think of a couple of approaches, but more information is needed.

Tom

[Mic35]

The SAS dataset is the same form as the one of the image (top). It’s the result of transpose and multiple merges.
The macro offers many possibilities (according to the settings) and requires an important data step at the end to show results...
It allows to obtain uni & bivariate results for continuous and categorical variables, to get RR/OR etc

i.e. %analysis(in=, out=, dep_var=, cont_vars=, class_vars=, silent=0, stddiff=2, RR=1, side=_vars_);

Mic

[PGStats · Accepted solution]

How about adding a data step such as:

data want;
set have;
array _field name_variable RR;
do i = 1 to dim(_field);
    if _field{i} = lag(_field{i}) then _field{i} = " ";
    end;
drop i;
run;

(untested)

[Mic35]

A big thank you to you PG.

Your code is very good.

With variables that appear in some cases and not in others (i.e. OR or RR), I added the code below to complete.
Now everything is perfect.
Thank you again.

ODS OUTPUT nLevels =work.nlevels;

PROC FREQ DATA= &out. NLEVELS ;
	TABLE _ALL_ / NOPRINT ;
RUN ;
PROC SQL NOPRINT ;
	SELECT STRIP(tableVar) INTO : listeDrop SEPARATED BY ","
	FROM  work.nlevesl
	WHERE NnonMissLevels = 0
	;
	ALTER TABLE  &out. DROP &listeDrop ;
QUIT ;