CODEWITHSUNDEEP
rdelimiterread.fwf
LillyA
LillyA

Reputation: 1

Is there a function in R to separate .txt files with no delimiter?

I have been given a dataset as a .txt format that I need to import in R for a research analysis. The .txt file has one column with numeric characters, no header, and no delimiter. These are a couple of the examples in the .txt file:

4878578572809275874037093859845083594859474905704627402739385785748756 0934893758795493758745846784678576857458708476968983984980985974687586 3989458476857609379087685796847586770493706759787398499485957658968590

I want to know how I can separate these characters in R according to its variables. So, I want to separate the first 3 numbers to be for variable 1, the next 5 numbers to be for variable 2, the next 2 numbers to be for variable 3, and so on.

Is there an expression I can use in R that will allow for me to separate the observations in the .txt file according to its required number of characters for each variable?

I have tried read.fwf but it did not work. Or maybe I tried it incorrectly.

Help with this issue will be much appreciated!

Upvotes: 0

Views: 1013

Answers (2)

David T
David T

Reputation: 2143

BETTER SOLUTION

library(utils)
raw <- "4878578572809275874037093859845083594859474905704627402739385785748756
0934893758795493758745846784678576857458708476968983984980985974687586
3989458476857609379087685796847586770493706759787398499485957658968590"

# Put your data in a temporary file.  You shouldn't have to do this, you data
# is already sitting in a file.
ff <- tempfile()
cat(file = ff, raw)

Now read it back in with read.fwf

answer <-  suppressWarnings(
  read.fwf(ff, widths = c(3, 5, 2)))
# Remember to clean up after ourselves.
unlink(ff)  # Again, you won't need to do this; your file isn't temporary.

answer    
   V1    V2 V3
1 487 85785 72
2  93 48937 58
3 398 94584 76

INITIAL ANSWER WITH REGEX

You can use regex (regular expressions). I've encoded the position breaks where you said to in your post:

library(tidyverse)
library(readr)
byRegx <- function(raw){
  rawSpl <- str_match(raw[1], "(?x) (^\\d{3}) (\\d{5}) (\\d{2}) (.+)")[1,]
  tibble(apples = rawSpl[2], bananas = rawSpl[3], carrots = rawSpl[4], 
         therestofthem = rawSpl[5])
}

Read your input to a table, then applied the byRegex function 

inputTbl<- tibble(
            raw = readr::read_lines("4878578572809275874037093859845083594859474905704627402739385785748756
                     0934893758795493758745846784678576857458708476968983984980985974687586
                     3989458476857609379087685796847586770493706759787398499485957658968590")) %>% 
  mutate(morecol = map(str_trim(raw), byRegx)) %>% 
  unnest() %>% 
  select(- raw)

inputTbl
# A tibble: 3 x 4
# apples bananas carrots therestofthem                                               
# <chr>  <chr>   <chr>   <chr>                                                       
# 1 487    85785   72      809275874037093859845083594859474905704627402739385785748756
# 2 093    48937   58      795493758745846784678576857458708476968983984980985974687586
# 3 398    94584   76      857609379087685796847586770493706759787398499485957658968590

Upvotes: 1

Mohanasundaram
Mohanasundaram

Reputation: 2949

You can use the substring function.

txt <- "487857857280927587403709385984508359485947490570462740273938578574875609348937587954937587458467846785768574587084769689839849809859746875863989458476857609379087685796847586770493706759787398499485957658968590"

select <-  c(3 , 5, 2)  # vector of number of words to be extracted

out <- as.numeric(substring(txt, 
                            cumsum(c(1, select[-length(select)])), 
                            cumsum(select)))
out
#[1]   487 85785    72

Upvotes: 0

Related Questions