Add comma to end of lines in text file

Johng · 08-21-2010, 03:22 AM

I have a plain text file with 360 lines of varying length text. How do I add a comma or other symbol to the end of each line so that I can convert the file to csv format that I can open in a spreadsheet (45 rows, 8 columns). That means each 8 lines of text forms 8 columns, with 45 rows.

druuna · 08-21-2010, 03:45 AM

Hi,

Adding a comma to the end of a line (all lines in a file):

sed -i.bak 's/$/,/' infile

This will make the changes in place, the original file will be saved with a .bak extension.

Hope this helps.

AnanthaP · 08-21-2010, 03:57 AM

Quote:

That means each 8 lines of text forms 8 columns, with 45 rows.

This actually makes me think that the OP has file where each row contains:

Row1Line1 Row1Line2 Row1Line3 Row1Line4 Row1Line5 Row1Line6 Row1Line7 Row1Line8
Row2Line1 Row2Line2 Row2Line3 Row2Line4 Row2Line5 Row2Line6 Row2Line7 Row2Line8
and so on.

If the OP clarifies what separates a "line" which is actually a column, then we can help further. For instance, if the separator is a blank space, then we would use sed to substitute blanks with commas.

OK

Johng · 08-21-2010, 04:08 AM

Thank you druuna and AnanthaP.

Adding the comma works, which is step one. However the spreadsheet still sees the modified file as only one column, which is what AnanthaP has indicated. So it appears that I now need to add lines 2,3,4,5,6,7,8 to the end of line 1, lines 10,11,12,13,14,15,16 to the end of line 9 etc, ending up with a file with only 45 lines?

J_Szucs · 08-21-2010, 04:16 AM

What is your field separator? A comma? Are commas found as field separator and not in fields?

If the answer is yes to all of the above questions, then this should work:
cat textfile.txt | sed 's/$/,/' | tr -d '\n' | sed 's/$\([^,]*,$\{8,8\}\)/\1\n/g' | sed -e 's/^ //' -e 's/$[ ]*,[ ]*$/,/g' -e 's/,$//g'

(first sed command adds comma at the end of each line, then tr command merges all lines, the second sed command breaks up the lines after each 8th commas, and the last sed command just doeas some "cleanup": deletes spaces from the beginning and end of rows and fields, and deletes comma from the end of each row)

konsolebox · 08-21-2010, 04:28 AM

in bash

Code:

#!/bin/bash

I=0

while read LINE;do
    (( ++I, I %= 8 ))

    if [[ I -eq 0 ]]; then
        echo "$LINE"
    else
        echo -n "$LINE,"
    fi
done

bash script.sh < input_file

Johng · 08-21-2010, 04:31 AM

Thank you J_Szucs, that did the job.

I was confused for a while - it wrote the modified output to the terminal, not modifying the original file.

grail · 08-21-2010, 04:32 AM

OP - generally, the way to stop all of us guessing and coming up with ways that don't actually help / work with your situation is to provide us
with a sample of the text file. Then we can actually see what unusual type characters there might be that need to be addressed in solving your problem.

J_Szucs · 08-21-2010, 04:40 AM

Sry, I do not really like overwriting the input file when testing a new function.
When it has proven to work, just add to the end of the command:
> textfile.txt.new; mv textfile.txt.new textfile.txt

grail · 08-21-2010, 05:15 AM

Assuming I am following:

Code:

awk '{line=line $0;f=1}!(NR % 8){print line;line="";f=0}f{line=line","}' oldfile > newfile