%EM_METACHANGE

[rmlmrmlm]

Hello, good evening.

Imagine a dataset named DSet with one variable, that contains variable names:

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

variable_name

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

income

interest

result

I want to transform this variables to level=ordinal, one by one.

So, I wrote this code inside an extended node.
I must say that this is inside a diagram with a strong legacy. I didn't build it but now I'm must maintain it.

As you can see m, I'm mixing macro code with names of dataset variable names ...

This isn't working ... possibilby due to this mess ...  name = variable_name

Can somebody help me, please?

%macro test;

DATA _NULL_;

            SET DSet;

            %EM_METACHANGE (name = variable_name, level = ordinal);

RUN;

%mend test;

%test;

[ballardw]

If you are wanting to call an already existing macro with values from a data set then the approach is likely to be CALL EXECUTE.

Perhaps:

DATA _NULL_;
   SET DSet;
   Call execute( '%EM_METACHANGE (name ='|| variable_name||', level = ordinal);');
RUN;

This is one time you don't want the macro inside double quotes because it will attempt to execute while the data step compiles.

This assumes that you have an existing macro %EM_METACHANGE that works correctly if you run it manually with the values of the variable.

There is not need for Macro Test that I can see.

[rogerjdeangelis · Accepted solution]

DATA _NULL_;
   SET DSet;
   call symputx('variable_name',variable_name);
   call symputx('ordinal',variable_name);
     rc=dosubl('
            %EM_METACHANGE (name = &variable_name, level = &ordinal);
     ');
RUN;