How to extract first few lines from a csv file inside a tar file without extracting it in linux?

Question

I have a tar file which has lot of csv files in it. How to get the first few lines of each csv file without extracting it?

I tried:

$(tar -Oxf $tarfile $file | head -n "$NL") >> cdn.log

But got error saying:

time(http:index: command not found

This is some line in one of the csv files. Similar errors are reported for all csv files... Any idea??

Answer 1

Using -O you can tell tar to extract a file to standard output instead of to file. So you should be able to first use tar tf <YOUR_FILE> to list the files from archive and filter it using grep to find the CSV files, and then for each file use tar xf <YOUR_FILE> <NAME_OF_CSV> -O | head tar xf <YOUR_FILE> <NAME_OF_CSV> -O | head to get the file's beginning to stdout. This may be a bit ineffective since you unpack the archive as many tiems as there are CSV files, but should work.

Answer 2

You can use perl and its Archive::Tar module. Here a one-liner that extract the first two lines of each one:

perl -MArchive::Tar -E '
    for (Archive::Tar->new(shift)->get_files) { 
        say (join qq|\n|, (split /\n/, $_->get_content, 3)[0..1]) 
    }
' file.tar

It assumes that the tar file only has text files and they are csv . Otherwise you will have to grep the list to filter those you want.