Finding sum per row based on conditions

[Anoush Khachatryan]

Members,

I have a dataset with multiple columns (var1, var2, var3, etc.) with missing observations. Per row, I want to find the sum of the values for var1, var2, var3, etc. that come after a missing value. For instance, if var1 is missing, I would only take the sum of var2, var3, etc. If var2 is missing, I would only take the sum of var3, var4, etc.

I would appreciate any assistance with this!

Thanks,
A

[William Lisowski]

What if var1 and var3 are both missing?

Also, please review the advice on providing example data given at #2 in your previous topic.

[Nick Cox]

The egen function rowtotal() ignores missings. If it is not what you want, I think we need a precise statement of how your rules differ from its rules, preferably with a data example showing what it does and what you want and how they differ.

[Anoush Khachatryan]

Thank you for your responses, William and Nick.

Here is the data I'm working with:

clear
input float(v1 v2 v3 v4 v5 v6)
.61 .29 . . . .12
.52 .63 . . . .11
.72 .79 .84 .43 .33 .
.57 .37 .22 . .71 .
end
tempfile dataset1
save `dataset1'

What I would like to do is add the values per row that come after a missing value. It should look something like this:

clear
input float(v1 v2 v3 v4 v5 v6 Total)
.61 .29 . . . .12 .12
.52 .63 . . . .11 .11
.72 .79 .84 .43 .33 . 3.11
.57 .37 .22 . .71 . .71
end
tempfile dataset1
save `dataset1'

If no values are missing, then I would add all values per row. If the last value is missing then I would add all values prior to that. If there is a jump (var4 is missing) then I would add only the values that come after var4. If var3 and var5 are missing, then I would only use var6.

I hope this clarifies my question a bit more. I would definitely appreciate any assistance!

A

[Nick Cox]

What a bizarre set of rules... Perhaps

egen wanted = rowtotal(v?)

replace wanted = max(var5, 0) + var6 if missing(var4)

replace wanted = var6 if missing(var3) & missing(var5)

[Anoush Khachatryan]

Thank you so much Nick! I really appreciate your help.