Every combination of categorical variables into new variable using IF THEN

[aowusudommey]

Hi all,

I'm having an issue that I hope you can help with. 

I have a very large dataset with multiple variables that I need to combine into various categories.

I've attempted this way:

IF (var1=1 and var2=1 and var3=1 and var4=1 and var5=1 and var6=1) OR

(var1=1 and var2=1 and var3=1 and var4=1 and var5=1 and var7=1) OR

(var1=1 and var2=1 and var3=1 and var4=1 and var5=1 and var8=1) OR

(var2=1 and var3=1 and var4=1 and var5=1 and var6=1 and var7=1) OR......

THEN combo6=1, else combo6=0

As you can see, this process has been tedious, and not very effective as the combinations i'm getting equal more than the data I have available. Does anyone have any idea how to accomplish this?

[PaigeMiller]

Forget SAS for a second. What is the logical rule that you are trying to program? Please explain.

[aowusudommey]

Sure. The logic rule I'm attempting to follow is pretty simple, but hard to describe in text. Essentially, we have A, B, C, D, E, F, G, and H. What I need is all of the combinations of these letters in sets of 8, 7, 6, 5, and 4. 

For example:

(A,B,C,D)=combo4

(B,C,D,E)=combo4

(C,D,E,F)=combo4

and so on. 

However, the dataset given to me is categorical. I only want the variables that are "yes" (A=1, B=1, etc), included in the new variable. 

8 is easy, but the remaining get complicated. Did that clarify at all?

[PaigeMiller]

---

@aowusudommey wrote:

Sure. The logic rule I'm attempting to follow is pretty simple, but hard to describe in text. Essentially, we have A, B, C, D, E, F, G, and H. What I need is all of the combinations of these letters in sets of 8, 7, 6, 5, and 4. 

For example:

(A,B,C,D)=combo4

(B,C,D,E)=combo4

(C,D,E,F)=combo4

and so on. 

However, the dataset given to me is categorical. I only want the variables that are "yes" (A=1, B=1, etc), included in the new variable. 

8 is easy, but the remaining get complicated. Did that clarify at all?

---

May we assume that if A is not equal to 1, then it must be zero?

[aowusudommey]

That is correct, each variable is assigned a value of either 1 or 0, there are no missing values. 

[aowusudommey]

So sorry, I should further explain that I would need any variable in the string that is equal 0 to cause the entire row to be equal to zero. 

For example

(A=1 and B=1 and C=1 and D=1), then combo4=1

However:

(A=1 and B=1 and C=1 and D=0) then combo4=0

[SASKiwi · Accepted solution]

Based on your incomplete explanation a statement like this may be helpful. 

combo6= (sum(of var1 - var8) = 6);

[aowusudommey]

I am unfortunately not looking for a sum, I'm looking to have every possible combination of these 8 binary variables to be assigned to a new variable. So that, for example, every combination of 6 of the 8 variables would equal combo6, every combination of 5 of the 8 variables would equal combo5, and so on. I'm very sorry, I am not able to give exact information without violating a privacy agreement. 

[SASKiwi]

But summing any combination of 8 binary variables where the total is 6 will give you one of your answers - combo6?

var1 = 1 plus var2 = 1 plus var3 = 1 plus var4 = 1 plus var5 = 1 plus var6 = 1 equals 6.

[aowusudommey]

You are absolutely correct, wow I cannot believe it was that simple. Thank you so much.