Split differents fixed width strings formats-CodePudding

Newbie question! I have a column with strings of two differents fixed widths formats. We can recognize the type of format by its name and split the string according to the format.

df <- data.frame(
var1 =  c('M1B123456789MM1158','M1C123456789zMM1183'),
var2 =  c('code1','code8'))

The fixed widths formats are:

formatM1B = c(3,9,2,4)
formatM1C = c(3,9,1,2,4)

So i hope this result:

 |format|var1_2   |var1_3|var1_5|var1_6|code |
1|M1B   |123456789|      |MM    |1158  |code1|
2|M1C   |123456789|z     |MM    |1183  |code8|

I tried the functions separate , str_split or str_split_fixed but i don't know how combine it with a sort of IF function to "test" or "regex" the format mentionned into the string. This question has certainly been asked a lot of time, i did hours research without being able to find something to adapt to my data :/

CodePudding user response：

library(tidyverse)

df %>% 
   extract(col= var1,into = c('format','1','2','3','4'), 
          regex = "^(M[1-9][A-Z])([1-9]{9})(z)?(M{2})([1-9]{4})")

The regex expresion has 5 groups:

(M[1-9][A-Z]): Search for a M, a int: 1,...,9, and an uppercase letter
([1-9]{9}): Search for 9 int numbers: 1,...,9
(z)?: Search if there is a z or skip
(M{2}): Search for 2 M
([1-9]{4}): Search for 4 int numbers: 1,...,9

Output:

  format         1 2  3    4  var2
1    M1B 123456789   MM 1158 code1
2    M1C 123456789 z MM 1183 code8

CodePudding user response：

Here is a function that does the splitting based on your formatM1B/C vectors,

f1 <- function(string, vec){
  start <- c(1, cumsum(vec)[-length(vec)]   1)
  end <- cumsum(vec)
  apply(data.frame(start, end), 1, function(i)substring(string, i[1], i[2]))
}

And we can apply it as,

Map(function(x, y) f1(x, y), df$var1,list(formatM1B, formatM1C))

#$M1B123456789MM1158
#[1] "M1B"       "123456789" "MM"        "1158"     

#$M1C123456789zMM1183
#[1] "M1C"       "123456789" "z"         "MM"        "1183"