how to find an exact substring match?

ldp · 02-21-2005, 01:36 PM

Hi,
I'm starting to learn about regular expressions and it all seems pretty nice to me but I'm missing something vital.
In my script, I want to find a substring in a certain string. It should return the position of the first character of this substring.

I tried the following:
fname="rancid_musicfile.wav"
pos=`expr index "$fname" wav$`; echo $pos => gives 2 as a position result but I'm expecting 18 as position because I want the complete substring to be taken in account, not just one of it's elements. (isn't it [wav] that matches any of the characters?)
I also tried different other things like:
pos=`expr index "$fname" "wav"$`
pos=`expr index "$fname" 'wav'$`
pos=`expr index "$fname" "wav$"`
pos=`expr index "$fname" 'wav$'`
But they all give the first occurence of the 'a' as a result.

Can someone point out the stupidity that I'm missing here? It must be something very obvious I suppose...

Thanks,
Lieven

ldp · 02-21-2005, 02:09 PM

I found a way to remove the wav extention from the string with the following: (But I would like to know it's position also.)
fname2=`echo $fname | sed -e 's/\.wav//'`
There might be easier ways to do this but it's only for learning purposes.

Hko · 02-21-2005, 02:36 PM

Quote:

Originally posted by ldp
I tried the following:
fname="rancid_musicfile.wav"
pos=`expr index "$fname" wav$`; echo $pos => gives 2 as a position result but I'm expecting 18 as position because I want the complete substring to be taken in account, not just one of it's elements. (isn't it [wav] that matches any of the characters?)

In shell wildcard matching "[wav]" means "any of the characters" (w, a or v in this case), but "expr" does not do do this for for entire strings, but only for single one characters in "wav".

ldp · 02-21-2005, 02:52 PM

So you're saying that with expr, I cannot look for the position of a substring but only for the position of a single character (or any one char from a list like "wav") in a string?

Hko · 02-21-2005, 04:13 PM

Quote:

Originally posted by ldp
So you're saying that with expr, I cannot look for the position of a substring but only for the position of a single character (or any one char from a list like "wav") in a string?

Yes.

Hko · 02-21-2005, 04:32 PM

Quote:

Originally posted by ldp
I found a way to remove the wav extention from the string with the following: (But I would like to know it's position also.)
fname2=`echo $fname | sed -e 's/\.wav//'`
There might be easier ways to do this but it's only for learning purposes.

Nothing wrong with yours, but a few other ways to do that:

Code:

fname2=`basename $fname .wav`

# or: (bash only, but no external programs used)
fname2=${fname%.wav}

To get the position of ".wav", you can now take just the length of fname2:

Code:

pos=`expr length $fname2`

# or: (bash only, but no external programs used)
pos=${#fname2}

As far as I know there's no utility that gets a substring index at once (a bit surprising I think). Though combining the two things gets close:

Code:

# Using $(..) instead of `..` which does the same thing,
# but $(..) can be nested (one inside the other). 
# This is bash-only though.

pos=$(expr length $(basename $fname .wav))

Or with some more heavy utility like "awk" you could do:

Code:

awk "BEGIN {print index(\"$fname\", \".wav\")}"

ldp · 02-22-2005, 01:34 AM

Thanks for your explanations.
Too bad there is no command/builtin or something to give the position of a substring. If I remember well, then most programming languages do have something like that.

kind regards,
Lieven

Hko · 02-22-2005, 06:28 AM

Yes, programming laguages generally do have some way to do that. I couldn't find a real simple way for scripts though.