Speedy/elegant way to unite many pairs of columns

Question

Is there an elegant/fastR way to combine all pairs of columns in a data.frame?

For example, using mapply() and paste() we can turn this data.fr

Answer 1

I'm not sure this is the best approach. See if the below code gives any speed improvement

require(dplyr)
transmute(mydf,x1=paste0( a.1,'.', a.2),x2=paste0( b.1,'.', b.2)) 

Answer updated based on comment :-)