extracting partial string 1

[tomte]

I want to extract all the characters after a certain string is found using fgrep, to the end of the current line.

fgrep 'abcde: ' filename

I want to extract to variable all the characters after the :

Thanks in advance

 

[CaKiwi]

You could do it all in awk

awk '{if (match($0,/abcde:/)) print substr($0,RSTART+RLENGTH+1)}' filename CaKiwi

 

[tomte]

Thank you I will give it a try.

 

[pavNell]

How 'bout...

while read LINE
do
myVar=`echo $LINE | sed 's/abcde:\(.*\)/\1/'`
done < filename > results

 

[PieterWinter]

How about

fgrep &quot;abcde:&quot; filename | while read line
do
echo ${line#*:}
done > results

The ${line#*:} will drop everything before and including the first :

 

[bi]

I have a slightly different question, but it's related to tomte's original question.

I need to do a find on a filename, then extract the path only, not the filename, from the results of the find. So, I would do a find on, say, passwd. The results are /usr/bin/passwd, /etc/passwd, and /etc/security/passwd. I want to get just /usr/bin, /etc, and /etc/security.

Thanks for any help.

 

[vgersh99]

#!/bin/ksh
a='/a/b/c/d'

echo &quot;a->[${a}] path2a->[${a%/*}]&quot;
vlad
+----------------------------+
| #include<disclaimer.h> |
+----------------------------+

 

[tomte]

Thanks to all who helped. I have a working extraction now, but I have an interesting situation. I have extracted the string to a variable NAME
I then want to append some string information to the string.
$NAME echos to the appropriate string ABCD
I am in a ksh script and I am using the syntax of
NAME=$NAMEyz
What I end up with is xyCD
I have tried NAME=$NAME'xy'
I have tried NAME=&quot;$NAMExy&quot;
and any other variation of this and end up with the same results. Any suggestions?

 

[SamBones]

Try...
[tt]
NAME=${NAME}xy
[/tt]

 

[tomte]

Actually, I have found that there is a ^M at the end of my string that is causing this to happen, but I cannot seem to sed out the character, anyone have a idea. I have tried sed 's/.^M//g' but it does not seem to take care of the issue.

 

[PieterWinter]

You could try

${NAME%^M}xy

Note that the ^M should be a &quot;control-M&quot;, not 2 separate characters.

 

[SamBones]

A lot of times, if you have an extra ^M at the end of your lines, it's because the file either came from a Windows/MSDOS machine, or was edited by a Windows/MSDOS editor (i.e. Notepad via SAMBA). The end of line character for Windows/MSDOS text files is a CR/LF pair, not just the LF that *nix uses. That extra CR is your ^M at the end of the line.

If this is the case, look for a utility called [tt]dos2unix[/tt]. It's purpose is to scrape those ^Ms off.

Hope this helps.