renaming the variable observations or combining the into one

[bondtk]

Hi

I have got data like this:

App_type_flag

Rural

Consumer

Commercial

Agri

Mining

Questions 1.    What I want to is : When I run the grouping of this data in proc sql...

                        I want to combine the Rural and Agri as one.

so the App_type_flag  should take Rural and Agri as one observation , 

so Rural + Agri  should be considered as Agri . So App_type_flag  just show,  Agri, consumer, commercial, mining .

What would be the appropriate way to do this:

Should I create a new variable or just change the rural to agri with some code,  please advise.

Question2 :

Also I want to segment the data to remove those records which are withdrawn and it matches the certain app_types

data tk.SBOS_final;

set tk.application_fj;

where final_decision_flag not in ('Withdrawn');

also want to add where app_type_flag not in ('TEMP' , ' ' )

so trying to say where app_type_flag doesn't match TEMP and blanks and the final_decision flag is not in Withdrawn,  how do I combine them and how to write the code effectively.

Please advise.

thanks

[novinosrin]

data have:

set have;

_App_type_flag=App_type_flag;

if App_type_flag= 'Rural' then  _App_type_flag='Agri';

run;

then do group by using _App_type_flag.. You can keep both the columns just to check the flow of data just in case.