R combine data frame columns with regex -

this question has answer here:

• reshaping multiple sets of measurement columns (wide format) single columns (long format) 5 answers

i have following data frame:

dat <- data.frame(     c = c(1 , 2) , a1 = c(1 , 2) , a2 = c(3 , 4) , b1 = c(5 , 6) , b2 = c(7 , 8) )   c a1 a2 b1 b2 1 1  1  3  5  7 2 2  2  4  6  8 

that merged columns based on shared prefixes become data frame:

dat2 <- data.frame(     c = c(1 , 2 , 1 , 2) , = c(1 , 2 , 3 , 4) , b = c(5 , 6 , 7 , 8) )   c b 1 1 1 5 2 2 2 6 3 1 3 7 4 2 4 8 

the way can think of try using melt(). attempt:

melt(dat , measure.vars = c(grep("^a" , colnames(dat)) , grep("^b" , colnames(dat))))     variable value 1 1       a1     1 2 2       a1     2 3 1       a2     3 4 2       a2     4 5 1       b1     5 6 2       b1     6 7 1       b2     7 8 2       b2     8 > 

needless say, incorrect.

we can use melt data.table takes multiple patterns in measure

library(data.table) melt(setdt(dat), measure=patterns("^a\\d+", "^b\\d+"),                     value.name=c("a", "b"))[, variable:= null][] #   c b #1: 1 1 5 #2: 2 2 6 #3: 1 3 7 #4: 2 4 8