Home / ubuntu / Questions / 1302367
Accepted
Cas
Asked: 2020-12-24 07:04:28 +0800 CST

How to grep series title from list of paths?

  • 772

I have the following command:

find /home/cas/plex-media/series/ -type f -name '*.srt' | grep -v .en.srt

Which will find all .srt files, that are not .en.srt in the directory and subdirectories of series.

This will give me a list that could look like this:

/home/cas/plex-media/series/Scorpion/Season 4/Scorpion - S04E06 - Queen Scary.srt
/home/cas/plex-media/series/Scorpion/Season 4/Scorpion - S04E03 - Grow a Deer, A Female Deer.srt
/home/cas/plex-media/series/Devs/Season 1/Devs - S01E03 - Episode 3.srt
/home/cas/plex-media/series/Modern Family/Season 8/Modern Family - S08E21 - Alone Time.srt

This is a simplified version. There could be 5 or 300 outputs, where there could be 8 or 50 times Scorpion for example. The output varies very much.

I trying to find a command (that I could pipe behind the original command), that will be give me the result:

Scorpion, Devs, Modern Family

And not:

Scorpion, Scorpion, Scorpion, Scorpion, Devs, Devs, Modern Family, Modern Family, Modern Family, 
Modern Family, Modern Family, Modern Family, etc..

You get the idea. For every show, one output. Not for every file.

It can be in another layout. It doesn't really matter (altough I'd prefer the layout above).

I've been trying around with grep but I just can't do it. So i'm basically asking: Do you have an idea how I could get the output above?

Thanks.

command-line

1 Answers

  1. Best Answer

    Assuming that the delimiter between the series title and the remainder of the filename is a consistent space-hyphen-space, you could do something like this with awk:

    find . -type f -name '*.srt' ! -name '*.en.srt' -printf '%f\0' | 
  awk -v RS='\0' -F ' - ' '!seen[$1]++{print $1}' | paste -sd,

    The !seen[$1] construct de-duplicates the results. The final paste command turns the output from a newline-delimited list to a comma-delimited list.

    • 1