data.frame of characters to vector-CodePudding

i have a data.frame containing X columns (i'll use three in this example):

data<- data.frame('col1'=c('gama','zeta','omega'),'col2'=c('square','circle','triangle'),'col3'=c('bike','car','airplane'))

i want to 'sum' the values of each row of this data.frame placing on them a separator, the values are characters. i want to do this without using a for loop (cause my real data.frame is very large).

the question drawing

the 'paste' function requires multiple vectors as input (when the input is a list of character the output is wrong), but the 'ncol' (number of columns) of the real data.frame will change all time (is a dynamic generated data.frame).

i dont know how to do this using the 'paste' or the 'stringr::str_c' functions, but i tried them and both return a wrong formated unique vector. and 'rowSums' function only accepts numeric (and dont accepts a separator, but i could handle this).

CodePudding user response：

You can use apply to apply the paste0 function on each row, and leverage the collapse parameter:

data$new_col = apply(data, 1, paste0, collapse=", ")

Output:

   col1     col2     col3                   new_col
1  gama   square     bike        gama, square, bike
2  zeta   circle      car         zeta, circle, car
3 omega triangle airplane omega, triangle, airplane

CodePudding user response：

We can use

library(tidyr)
unite(data, new_col, everything(), sep = ", ", na.rm = TRUE, remove = FALSE)

-output

                    new_col  col1     col2     col3
1        gama, square, bike  gama   square     bike
2         zeta, circle, car  zeta   circle      car
3 omega, triangle, airplane omega triangle airplane