topic Re: Script to select text from a line in Operating System - HP-UX

Script to select text from a line

Luk Vandenbussche — Mon, 14 May 2007 01:36:02 GMT

Hi,

I have the following question.
I have a file with several lines in it.
In each line a have the adress off a website, not on a fixed position. I need a script / command to filter out this website.

fe
The file contains

aeratarawww.hp.comqmlkmlk
qdfdwww.google.commmkkm
arewww.gmail.com kmkùkù

I need as output only

www.hp.com
www.google.com
www.gmail.com

Please advise

Re: Script to select text from a line

Masatake Hanayama — Mon, 14 May 2007 01:57:43 GMT

Below is not perfect, but may work for the example.

sed -e 's/^.*www\./www\./' -e 's/\.com.*$/\.com/' filename

Re: Script to select text from a line

Dennis Handly — Mon, 14 May 2007 02:11:55 GMT

If they start with www. and end in .com you can use:
sed -e 's/*.$www\..*\.com$.*$/\1/' filename

Re: Script to select text from a line

john korterman — Mon, 14 May 2007 04:01:08 GMT

Hi Luk,

try something like this, using your file as $1:

#!/usr/bin/sh
while read line
do
STRING1=$(echo ${line%%www\.*})
STRING2=$(echo "${line#$STRING1}" )
STRING3=$(echo ${STRING2##*\.com})
echo ${STRING2%$STRING3}
done <$1

regards,
John K.

Re: Script to select text from a line

Rasheed Tamton — Mon, 14 May 2007 04:11:10 GMT

Hi,

# awk 'BEGIN {FS="."} {printf "www."$2".com\n"}' filename
www.hp.com
www.google.com
www.gmail.com

Re: Script to select text from a line

Rasheed Tamton — Mon, 14 May 2007 05:13:13 GMT

Hi,

With more error checking.

awk -F"." '/www\./ && /\.com/ {print "www."$2".com"}' filename

If the file will definitely contain the strings as "www.xxx.com" form, then here is a more simplified form:

awk -F"." '{print "www."$2".com"}' filename

Re: Script to select text from a line

Hein van den Heuvel — Mon, 14 May 2007 07:11:30 GMT

Luk, what problem are you really trying to solve? If the file comes from a particular file (index.dat) or log then others may have dealt with that before.

I find it unlikely that the www just starts after random text and the .com runs into random text as your example suggests. Is there really no seperator of sorts (whitespace?) If there a field lenght/offset indicator elsewhere in the structure?

If the text is all we have, is the www and .com, and a single field between them a complete description, or does the code also have to trigger on .org, .net and what have you not?

Cheers,
Hein.

Re: Script to select text from a line

spex — Mon, 14 May 2007 08:10:35 GMT

Hi Luk,

This sed script is a bit more generic than some of the other solutions:

$ cat urls
aeratarawww.hp.comqmlkmlk
qdfdwww.google.commmkkm
arewww.gmail.com kmkyky
jkddswww.hello.orgafsdjk
wjekwjws1.hp.netfds7u8
weewwemail.dsdsdswww.netsdfjkldskdsj
wwewuiweweb.ieee.orgfdjkdjdk
www.ibm.com

$ sed -e 's/^.*$...\..*\....$.*$/\1/' urls
www.hp.com
www.google.com
www.gmail.com
www.hello.org
ws1.hp.net
ail.dsdsdswww.net
web.ieee.org
www.ibm.com

PCS

Re: Script to select text from a line

Luk Vandenbussche — Mon, 14 May 2007 09:00:03 GMT

Thanks for the help