R collapse multiple rows into 1 row - same columns

Question

This is piggy backing on a question I answered last night as I am reconsidering how I\'d like to format my data. I did search but couldn\'t find up with any applicable answe

Answer 1

You can reshape to long format, drop the blank entries and then go back to wide:

res <- dcast(melt(df, id.vars = "record_numb")[ value != "" ], record_numb ~ variable)

   record_numb col_a col_b col_c
1:           1   123   234   543
2:           2   987   765   543

You may find it more readable at first using magrittr:

library(magrittr)
res = df %>% 
  melt(id.vars = "record_numb") %>% 
  .[ value != "" ] %>% 
  dcast(record_numb ~ variable)

The numbers are still formatted as strings, but you can convert them with...

cols = setdiff(names(res), "record_numb")
res[, (cols) := lapply(.SD, type.convert), .SDcols = cols]

Type conversion will change each column to whatever class it looks like it should be (numeric, integer, whatever). See ?type.convert.