Ytdyklly

I have a large csv file which contains multiple columns. I want the spaces to be removed from the first column and want to create that as a new column.

Example...

Input:

a b,xyz,d e f    

a b c,xyz,d e f    

a b c d,xyz,d e f   

Output:

ab,a b,xyz,d e f   

abc,a b c,xyz,d e f   

abcd,a b c d,xyz,d e f    

Using sed:

sed 'h;s/,.*/,/;s/ //g;G;s/n//' data.csv

Explained:

h - stash current line to the hold space

s/,.*/,/ - remove everything after the first comma

s/ //g - remove spaces

G - append the line from the hold space back to the pattern space

s/n// - remove extra newline, as left by G

Using bash with standard tools:

$ paste -d ',' <( cut -d ',' -f 1 file | tr -d ' ' ) file

ab,a b,xyz,d e f

abc,a b c,xyz,d e f

abcd,a b c d,xyz,d e f

paste combines the input from the two given files with a comma as the delimiter.

The first file is produce by a process substitution which extracts the first comma-delimited column of the file and removes its spaces.

The second file is simply the original unmodified file.

bash solution assuming a file test.txt:

#!/bin/bash



while read -r line; do

   IFS=',' read -ra fields <<<"$line"

   (IFS=','; echo "${fields[0]// /},${fields[*]}")

done <"test.txt"



exit

This takes advantage of IFS to read the values of the csv into an array and output them. Read each line of test.txt into a string, then read this line into an array by splitting on commas. Print the first element of this array with spaces removed, then the entire array. A quoted array using * will output the elements separated by IFS, which we previously set to a comma. The inline declaration of IFS in the read statement and inside a subshell () will preserve the value of IFS for the rest of the script's execution.

awk 'BEGIN { FS=","; OFS="," } { col=$1; gsub(/ /, "", col); print col,$0 }' file

Use commas as input and output column separators. Extract first column and sub out spaces. Output the modified first column followed by the original line.

with the great Miller is:


mlr --csv --implicit-csv-header --headerless-csv-output
put '$newField=gsub($1," +","")'
then reorder -f newField input.csv


$1 is the first field. Than I apply a regex search and replace to obtain the new field.