CODEWITHSUNDEEP
rlistvectordata.table
frankowiczm
frankowiczm

Reputation: 13

R data.table: sum of vectors in seperate columns

I have a data.table with vectors or actually list with vector inside columns cells i.e.:

testDATA <- data.table('COLUMN_1' = list(1:4, 11:14, 21:24),
                       'COLUMN_ABC' = list(5:8, 25:28, 15:18),
                       'COLUMN1342' = list(2:5, 32:35, 22:25))

Additionally I have a vector with names of columns which vectors' elements I want to sum i.e.:

columnsToSum <- c('COLUMN_ABC', 'COLUMN1342')

How can I add column to data.table which will sum vectors' elements from columns defined in columnsToSum so that in the first row I will have vector c(6,8,10,12) ? I know that I can sum two vectors like this:

testDATA[, sumCol := list(Map(function(x,y) x + y, get('COLUMN_ABC'), get('COLUMN1342')))]

although the thing is that I have all columns which I should sum in columnsToSum vector. Any ideas?

Upvotes: 0

Views: 593

Answers (1)

MrFlick
MrFlick

Reputation: 206308

I think this should work for you

testDATA[, sumCol := Reduce(function(a,b) Map(`+`,a,b), .SD), 
    .SDcols=columnsToSum]

We use .SD to get a group of columns from the data table, then we use Reduce() to sequentially add them up and use Map() to add up each of the rows separately. It will work with more than 2 columns

columnsToSum <- names(testDATA)
testDATA[, sumCol := Reduce(function(a,b) Map(`+`,a,b), .SD), .SDcols=columnsToSum]
testDATA
#       COLUMN_1  COLUMN_ABC  COLUMN1342          sumCol
# 1:     1,2,3,4     5,6,7,8     2,3,4,5     23,31,39,47
# 2: 11,12,13,14 25,26,27,28 32,33,34,35 193,201,209,217
# 3: 21,22,23,24 15,16,17,18 22,23,24,25 153,161,169,177

Upvotes: 2

Related Questions