Create a new variable based on two others

[ramgouveia]

I have this dataset

 and I want to create a new variable based on C3 and C3_F.

When C3 has a value I want this value. When not, like the green example shown, I want the C3_F value.

Can you help me please?

[Patrick]

You could use the coalescec() function documented here. 

[ramgouveia]

Patrick

The link doesn't work.

[Patrick · Accepted solution]

The link works for me.

Here how this could look like in a data step. Make sure to define the length of new_var long enough to hold the string (=the greater length of your two source variables)

length new_var $20;
new_var=coalescec(cae3,cae3_fue);

[andreas_lds]

@Patrick the link does not work:

Whitelabel Error Page

This application has no explicit mapping for /error, so you are seeing this as a fallback.

Wed Apr 07 12:26:11 UTC 2021

There was an unexpected error (type=Not Found, status=404).

[ramgouveia]

Thank you very much @Patrick !

It worked!

I used this code:

DATA D;
set D;
length C3_E $20.;
C3_E=coalescec(c3,c3_f);
run;

[Patrick]

Don't just use the length of $20 from my sample but chose a length which matches your source variables. If the length is too short then you will truncate the string, if it's too long then you're wasting space.

You can run below script. Then look-up in the report what lengths your two source variables have and use the bigger of the two values in the length statement for your new variable.

proc contents data=DEM_MES.MERGEEFATURADMRSS202001;
run;quit;

[Kurt_Bremser]

I took the liberty to update the link, it should work now everywhere.

(I just copy/pasted the link that I get from the documentation website, I have no clue what was wrong)

[Patrick]

Thank you.