0

I have rows and columns preceded by an introduction that I do not need. There are no headers for the columns. The data in my current file looks something like this (IP Addresses are fake):

This is a totally extraneous introduction and does not have anything to do with the data. It is here as a facsimile of what the output file looks like.

df    bank.com 10.10.10.1
sdfdg store.com 10.10.10.2
s     church.com 10.10.10.3

I need to skip past the introduction, paste in the extracted data to look like this below (derived from field two above), and put it all into a new .txt file. The strings do not need quotes:

bank.com,store.com,church.com

Any advice on how to do this in Bash?

I tried using this following technique, but it only grabbed the first line of the introduction, and did not go through each row.

Turning multi-line string into single comma-separated

Carl Gustav
  • 9
  • 2
  • You want to skip the first two lines? – Cyrus Apr 01 '21 at 22:15
  • I need to skip the introduction and grab data starting with the first row. I only need the second element from each row, saving those elements in a new file where they are separated by a comma. – Carl Gustav Apr 01 '21 at 22:21

2 Answers2

1

Assumptions:

  • always skip the first 2 lines of the input file
  • the second field contains no white space
  • there are no blank lines

Sample data file:

$ cat input.dat
This is a totally extraneous introduction and does not have anything to do with the data. It is here as a facsimile of what the output file looks like.

df    bank.com 10.10.10.1
sdfdg store.com 10.10.10.2
s     church.com 10.10.10.

One awk solution:

$ awk 'FNR>2 {printf "%s%s", pfx, $2; pfx=","} END {printf "\n"}' input.dat
bank.com,store.com,church.com

Explanation:

  • FNR>2 - for record (row) numbers greater than 2 ...
  • printf "%s%s", pfx, $2 - print our prefix (initially blank) plus field #2; because there is no \n in the format the cursor is left on the current line
  • pfx="," - set prefix to a comma (,) for the rest of the file
  • END {printf "\n"} - add a \n to the end of the line
markp-fuso
  • 28,790
  • 4
  • 16
  • 36
0

This should work:

tail -n +3 filename | awk '{print $2}' | sed -z 's/\n/,/g;s/,$/\n/' > newfile.txt

The tail -n +3 will skip the first 2 lines of the file (so change this to however many lines the intro is, plus one).

The next part prints out only the second column, the one you're interested in.

The third replaces the newlines with commas.

The last part places the output into a new file.

Eric Miller
  • 56
  • 6