How to add two columns in a data frame to create a new third column, based on sub-strings of column names, in R?

Question

Lets consider a simple data frame as follows :

id area1feature1 area1feature2 area2feature1 area2feature2
1  1             2             3             4
2  3             6             1             5

Now I would like to combine feature1 for all areas, feature2 for all areas and so on, and then create a new sumOfFeature1 , sumOfFeature2 , etc.

So the expected output is something like this :

id area1feature1 area1feature2 area2feature1 area2feature2 sumOfFeature1 sumOfFeature2
1  1             2             3             4             4             6
2  3             6             1             5             4             11

How can I match columns based on sub-string and then combine them to create new columns for data frame?

Answer 1

The way I did it is as follows : Let input be the data frame.

features_to_be_combined <- c('feature1', 'feature2')
locations <- sapply(features_to_be_combined, grep, colnames(input))
feature1_locations <- locations[, 'feature1']
sumOfFeature1 <- rep(0, dim(input)[1])
for (i in 1:length(feature1_locations)) {
    sumOfFeature1 <- sumOfFeature1 + input[, feature1_locations[i]]
}

Now all that remains is to repeat the same procedure for feature2 and then add newly created features, namely sumOfFeature1 and sumOfFeature2 , to the input data frame. I am sure there will a better way to do this (may be using apply again on combined features), but this worked for me as expected.