Copy .TS files but not .2.TS files?

redpenguin2 · 07-19-2019, 12:47 PM

I have a former Mythbuntu/Ubuntu system running MythTV.

My MythTV outputs .ts files and I have a script called Mythlink.pl.

MythTV just outputs weird names like 7006_2343644.ts so the script makes a folder of sym links named like "ShowName-Title-RecordDate.ts".

The problem is, for many files lately it has been creating two sym links.

A.ts & A.2.ts.

Because they are sym links, I have to use the cp command to copy them to another location.

My problem is, how can I easily tell cp, I only want .ts not .2.ts? I know I can't just use ShowName*.ts as that picks up the .2.ts files also, then I have to delete all the extra afterwards.

273 · 07-19-2019, 12:51 PM

Does find let you specify the end of a line? Then pipe it to cp?

redpenguin2 · 07-19-2019, 12:56 PM

Quote:

Originally Posted by 273

Does find let you specify the end of a line? Then pipe it to cp?

Yes, but I am not quite sure what to feed to find as something like *.ts just gives the .2.ts files also.

scasey · 07-19-2019, 01:19 PM

Two thoughts:
Delete the extras before you copy:

Code:

rm *2.ts

-- that would also work afterwards.

Address the glitch in Mythlink.pl that's creating the 2.ts files in the first place.

273 · 07-19-2019, 01:26 PM

Quote:

Originally Posted by redpenguin2

Yes, but I am not quite sure what to feed to find as something like *.ts just gives the .2.ts files also.

i am suggesting you man find and look for a way to specify end of line.

273 · 07-19-2019, 01:29 PM

Hmm, perhaps it's not that simple:
https://unix.stackexchange.com/quest...-line-with-sed

BW-userx · 07-19-2019, 01:42 PM

give this a go.

Code:

#!/bin/bash

while read f
do
	g=${f//./ }
	[[ ! "$g" =~ '2' ]] && echo $g
done < <(find $1 -type f -iname "*.ts")

removes the . (period), hunts for the 2, if not 2 within string, then act on it, or

Code:

#!/bin/bash

while read f
do
	g=${f//./ }
	[[ ! "$g" =~ '2' ]] && { echo "$f" ; } || { rm "$f" ; } 
done < <(find $1 -type f -iname "*.ts")

###################

#!/bin/bash
copy2=/someWhereElse
while read f
do
	g=${f//./ }
    if	[[ "$g" =~ '2' ]] ; then
          rm "$f"
     else 
         cp "$f" "$copy2"
    fi
done < <(find $1 -type f -iname "*.ts")

scasey · 07-19-2019, 02:12 PM

Code:

[[ "$g" =~ '2' ]]

[[ ! "$g" =~ '2' ]]

Too cool! Looks just like perl...did not know that would work in bash!

redpenguin2 · 07-19-2019, 02:45 PM

Wow thanks folks for all your replies!

I will have to study the examples further.

I know basic wildcards but have been having trouble trying to learn RegEx.

I have done text replacement with sed, tr, gawk, etc so I know the general concept just have to get the regex patterns down.

jefro · 07-19-2019, 02:58 PM

Some of the members at LQ are awesome with this stuff as you can see.

scasey · 07-19-2019, 03:08 PM

The regex BW-userx has suggested are testing for the presence (or absence) of a 2 in the file name.
One can use globbing to test for presence:

Code:

*2.ts
or just
*2*

but I'm not thinking of a way to use globs to test for the absence of a 2.

Don't confuse the two. Shell globbing is not regex. Different syntax and capability entirely. Some of the same symbols, tho, just to be confusing.
Your original question is about globbing. (which is not to say that a regex solution is in any way wrong...I certainly liked what BW-userx posted!)