Insert a comma after a fixed number of characters or a space-CodePudding

I'm wondering if there is an option to insert an "or" within sed, I need to convert a txt file to a csv file, the thing is my first column values can be up to 30 characters (minimun 1 character) if this happens the value in the next column start after the last character, otherwise space(s) are added, therefore I need to validate both cases before insert a comma, otherwise the comma might be inserted in the incorrect place.

Example:

column 1                      column 2
column1-value-up-to-30-charssscolumn2-value
column1-second-value          column2-second-value
column1-third-value-value     column2-thrid-value-value

Expected result (comma inserted after each value, even column names)

column 1,column 2
column1-value-up-to-30-charsss,column2-value
column1-second-value,column2-second-value
column1-third-value-value,column2-thrid-value-value

Therefore, I cannot always used sed to look for and empty space (like in value one) because the comma will be inserted at the end of the value in column 2

CodePudding user response：

Tested with GNU awk:

awk 'BEGIN{FIELDWIDTHS="30 30"; OFS=","}
     {
       sub(/  $/, "", $1);
       sub(/  $/, "", $2);
       $1=$1;
       print
     }' file

FIELDWIDTHS: A whitespace-separated list of field widths

Output:

column 1,column 2
column1-value-up-to-30-charsss,column2-value
column1-second-value,column2-second-value
column1-third-value-value,column2-thrid-value-value

See: 8 Powerful Awk Built-in Variables – FS, OFS, RS, ORS, NR, NF, FILENAME, FNR

CodePudding user response：

With your given input, maybe something like this?

sed -E 's/(.{1,30})(.*)$/\1,\2/;s/[[:space:]]{2,}//' file.txt