Whether the next calculation cover 100% the previous calculation?

[ResoluteCarbon]

Hi all,

When calculating one variable, I accidentally calculate the same named variable by two different methods, they all in the same dataset:

x= y-z;
x=m-n;

I am wondering whether the previous x will be replaced totally by the later x? Or just be replaced partly? Is there any exception that the later x just replace the former x partly?

Thank you.

[ballardw · Accepted solution]

---

@ResoluteCarbon wrote:

Hi all,

When calculating one variable, I accidentally calculate the same named variable by two different methods, they all in the same dataset:

x= y-z;
x=m-n;

I am wondering whether the previous x will be replaced totally by the later x? Or just be replaced partly? Is there any exception that the later x just replace the former x partly?

Thank you.

---

Yep.

When executing code in a data step it evaluates things from the "top" of the data step code to the bottom.

So if you assign a different value on a line below the first assignment the bottom one is the result.

The "exceptions" would all come if one or both of the assignments are conditional.

If sex='F' then x = (some calculation based on the sex);

else if sex='M' then x = (different calculation based on sex);

There is no "partly" unless you mean a character value where you assign different bits and pieces, which is possible but you would need to do the coding to implement such. With numeric values, not going to happen.