UNIX for Dummies Questions & Answers

Hi,

I want to extract a part of filename and pass it as a parameter to one of the scripts. Could someone help.

File name:-

NLL_NAM_XXXXX.XXXXXXX_1_1.txt.

Here i have to extract only XXXXX.XXXXXXX and the position will be constant. that means that i have to extract some n characters from xth position.

Thanks.

Code:

f=NLL_NAM_XXXXX.XXXXXXX_1_1.txt
s=`echo $f | cut -c 9-21`
echo "$f -> $s"

Code:

echo "NLL_NAM_XXXXX.XXXXXXX_1_1.txt" | sed 's/_[0-9].*.*$//;s/^.*_//'

or in BASH:

filename=NLL_NAM_XXXXX.XXXXXXX_1_1.txt

# ${stringosition:length}
extract=${$filename:9:13} # nb. no spaces around the colons

echo $extract

see http://tldp.org/LDP/abs/html/refcards.html#AEN20417
in the advanced bash scripting guide.

Thank you all.

I just used this code and it worked:-

f=NLL_NAM_XXXXX.XXXXXXX_1_1.txt
s=`echo $f | cut -c 9-21`
echo "$f -> $s"

The solutions that have been suggested will work fine, so long as the length of the filename does not vary. If it does, then you need a different solution.

One possibility is this:

f=NLL_NAM_XXXXX.XXXXXXX_1_1.txt
f1=${f%%.*} # Extracts part of name before first period (NLL_NAM_XXXXX)
f2=${f#*.} # Extracts part of name after first period (XXXXXXX_1_1_.txt)
fileName=${f1##*_}.${f2%%_*}

The generation of fileName deserves a little explanation. What I have done is split the filename into the piece before the first period and the piece after. So the first piece (f1) contains the name up to XXXXX and the second piece (f2) contains the piece from XXXXXXX on. Next I strip the part of f1 up to and including the last underscore (${f1##*_}, leaving XXXXX, concatenate with a period (.) and what is left of f2 after removing everything from the first underscore on (${f2%%_*}. The resulting string (XXXXX.XXXXXXX) is assigned to the variable fileName.

This solution will work in place of the other solutions that rely on extracting a fixed length segment of srting starting a a fixed position. The difference is that the dependence on length is eliminated. The only assumption made about the string is that there is no period before the XXXXX.XXXXXXX in the original filename.

Although Mark's is an elegant solution, there's a simpler way to do and still remain independent of filename length: use cut's field delimiter option:

f=NLL_NAM_XXXXX.XXXXXXX_1_1.txt
s=`echo $f | cut -d "_" -f3`
echo "$f -> $s"

Here, we tell cut to use underscore ("_") as the field delimiter. cut now sees the string as five fields (NLL, NAM, XXXXX.XXXXXXX, 1 and 1.txt respectively) and spits out the third.