shell script problem..

[nomanoma]

Helloo..

I have no idea where to start with this problem...

In my directory there are ~1000 files..

each ones name starts with number..

for example..

12file1
13file124
142file5
12file34
13file9
1file12

ok I sorted them but how to merge thoese with same digits numberst in their name..

to have just

1file
12file (12file1 + 12file34)
13file (13file124 + 13file9)
142file


thanks..any help in appretiated..

 

[kasparov]

This will kind of work - but it doesn't sort the output properly ... I'm sure others will have better ways to do it.

sed 's/file/&]/' < data | sed 's/].*//' | sort | uniq

Chris

 

[kasparov]

Sorry - my file containing the input data is called data.

 

[kasparov]

Sorry (again) I didn't read your question properly - I think this is what you want?

for file in `cat filenames`
do
SHORT_NAME=`echo $file | sed 's/file/&]/' | sed 's/].*//'`
cat $file >> $SHORT_NAME
done

(my list of filenames is now called filenames)

Chris

 

[nomanoma]

Thanks...but I am not quiet sure for the meaning of

SHORT_NAME=`echo $file | sed 's/file/&]/' | sed 's/].*//'`
cat $file >> $SHORT_NAME


Also I hope that content of 12file will be 2 files

(12file1 + 12file34) etc.
can U just help me out with meaning of whose 2 sed s...everything elese is clear to me..

Thanks..

 

[kasparov]

Hi

Probably the easiest way to figure out the code is run it at the command line bit by bit. (You'll have to use a real filename, not $file

echo $file ; echoes the file name

echo $file | sed 's/file/&]/' ; puts a ']' after the string 'file' in the filename - so 12file123 becomes 12file]123

echo $file | sed 's/file/&]/' | sed 's/].*//'` ; removes the ']' and the rest of the filename - so 12file]123 becomes 12file

then all files which begin 12file are put into a file called 12file. (This assumes that none of your files has a ']' in the filename)

Was this what you wanted?. If you test it by creating a few small files with names like the examples you gave hopefully you'll see it does what you want. (I've tested this but only on a small set of files - I'm assuming it works in all cases, I think it should)

Let me know if it's OK, Chris

 

[PHV]

sed 's/file/&]/' | sed 's/].*//'
Why not simply this ?
sed 's!file.*!file!'


Hope This Helps, PH.
Want to get great answers to your Tek-Tips questions? Have a look at FAQ219-2884 or FAQ181-2886

 

[kasparov]

Why not indeed?

Thanks PHV

 

[nomanoma]

Hmm...what to say..It is working..I do appreciate your help...

And thanks for explanation..but this PHV's code is not clear to me..



Thanks it works

cheers