COnverting missing variables to 0

[sree0203]

Hi,

To convert categorical variables to numbers I have used the following code in my program

If brand="minute.maid" then mid=1;
Else if brand =^"minte.maid" then mid =0;

But in the output for mid variable all the fileds with 1 are appearing and fields with 0 are appearing missing values.

Inorder to evade this I have used options missing ="0"

But while I am proc reg funtion the zero values are still getting considered as missing variables only.

How to get rid of this problem?

(Sent from handheld device please bear with typos)

[Astounding]

You have the characters in reverse order.  To get "not equal" you can use:

ne

^=

But not:

=^

[Reeza]

---

@sree0203 wrote:
Hi,




But in the output for mid variable all the fileds with 1 are appearing and fields with 0 are appearing missing values.

Inorder to evade this I have used options missing ="0"

---

No, it means you have missing values not 0, which indicates something is wrong with your code. Fix your IF/THEN statement as indicated by @Astoundingso you get 0's not missing values.  

[RW9]

Simplfy your if statement and you will get the correct answer.  Binary choices are best mae using ifn/ifc:

mid=ifn(brand="minute.maid",1,0);

[ballardw]

Perhaps

Else if missing( brand)  then mid =0;

If you have many of these types of operations to do you might consider a custom informat.

proc format library=work;
invalue missingtext
. = 0
other=1;
run;

data example;
   x= ' ';
   yx=input(x,missingtext.);
   z='anything';
   yz=input(z,missingtext.);
run;

This informat looks at an input string, if it is missing then the result is 0 anything else is 1. Even if the input string has a leading blank the result is one.

For your use then you would have

mid=input(brand,missingtext.);

NOTE this works if you are interested in knowing that your variable has some text.