Hello,

I have data that looks like the following

Code:
* Example generated by -dataex-. For more info, type help dataex
clear
input double(it_budget100_657 it_budget101_1267 it_budget102_730 it_budget103_1000 it_budget104_64)
 56810   .      . 28136             3475
131975   .      . 14813            12850
367000   .      . 29000   22914.84765625
331000   .      . 24000     39438.359375
276000   .      . 21000  30441.775390625
105000   .      . 13000 11991.0087890625
131300 129 151500     .            20285
241794   . 275696     .            34261
278940   . 276936     .            36480
277691   . 275696     .           275696
276178   . 274194     .           274194
253459   . 274194     .                .
end
What I am trying to do is loop through each variable name and extract the first number which I would then save as a local. Then, within this loop, I would loop again through the variable names and extract the first number where I check to see if this extracted number is the same as the one saved in the outer loop.

For example, in the first loop I would extract the number 100 somehow and then save this as a local. Within this loop, I would extract the next number 101 and then check to see if it is the same as 100. If so, I would use pwcorr to calculate correlations between the two vectors.

For context, the first number corresponds to a bank (like Wells Fargo), and each variable is a unique branch-bank pair (so some Wells Fargo in the United States). What I want to do is calculate correlations for branches that are not in the same bank. Hence, this is why I would like to skip the branches that have the same first number. I am really only struggling with how to extract the first number from each variable name. Thanks for any help!