pair_blocking can use vector of columns with and / or

[hswerdfe]

pair_blocking currently takes a vector of columns and all items must be the same to generate a potential pair.

p <- 
  noisy %>% 
  select(-key) %>% 
  colnames() %>% 
  map_dfr(~pair_blocking(noisy, noisy, blocking_var=.x, large = F))

on my testing data as leaving it unblocked , generates to many pairs for my machine but no one column seems sufficient to generate a block as any blocked column could have a spelling mistake.

FYI : I am using reclin from CRAN