Remove everything before... in a string

I

IT_Architect

I read a string in a file with the command you see below
[cmd=]egrep -i "Volume.Capacity_datastore1:" /tmp/vmware_api_10.17.127.132.txt[/cmd]
The string it returns is:
Code: 
Volume.Capacity_datastore1:     122406567936
Then I added this:
[cmd=]egrep -i "Volume.Capacity_datastore1:" /tmp/vmware_api_10.17.127.132.txt | sed 's/[[:space:]]//g'[/cmd]
and now the string it returns is:
Code: 
Volume.Capacity_datastore1:122406567936

How can I remove from the beginning of the line through the ":"?

Edit: I guess I had to post here to figure it out. The answer is:
[cmd=]egrep -i "Volume.Capacity_datastore1:" /tmp/vmware_api_10.17.127.132.txt | sed 's/[[:space:]]//g' | sed 's/^.*://g'[/cmd]
 
IT_Architect said:
How can I remove from the beginning of the line through the ":"?

Edit: I guess I had to post here to figure it out. The answer is:
[cmd=]egrep -i "Volume.Capacity_datastore1:" /tmp/vmware_api_10.17.127.132.txt | sed 's/[[:space:]]//g' | sed 's/^.*://g'[/cmd]
Click to expand...

Removing the spaces separately is not necessary. .* matches any character, so the second sed(1) command that deletes everything from the start of the string to the colon is all that's needed. Beware of regular expression "greediness":
Code: 
% echo "   abcdef:123:ghi" | sed 's/^.*://g'
ghi

The regex is greedy and takes the biggest match it can find.
 
cut() can also work. as with wblocks echo statement:

Code: 
% echo "   abcdef:123:ghi" | cut -d ':' -f3
ghi

Both ways are appropriate. Even using tr() and awk() could work. The cut command would take the least memory out of all the solutions.
 
Both replies were a HUGE help. They both helped me get a handle on working with streams, sed and cut. One good example is worth a 1000 man pages. LOL! I couldn't use them only because they didn't trim the leading spaces in front of 122406567936, but they helped me in my next hurdle tremendously.

Thanks!
 
If you know for sure that there will be only spaces and not other whitespace like a tab in front of the numbers, sed(1) can handle that:
% echo "Volume.Capacity_datastore1: 122406567936" | sed -e 's/^.*:\( \)*//'

If there could be tabs, Perl's regex handling is far, far more powerful than sed(1)'s 1972 view of the world:
% echo "Volume.Capacity_datastore1: 122406567936" | perl -pe 's/^.*:(\s)*//'

\s means whitespace, including tabs.
 
wblock@ said:
If you know for sure that there will be only spaces and not other whitespace like a tab in front of the numbers...If there could be tabs, Perl's regex handling is far, far more powerful
Click to expand...
"When the only tool you know is a hammer, everything looks like a nail." I'm VERY open to your suggestion because, as you already know, I'm at least as likely to get tabs as spaces in the output of a machine. I know PHP but nothing about Perl. I'm not great with regex, but I know enough to convert POSIX to PCRE to make PHP scripts compatible for PHP 5.3+.

Thanks!
 
wblock@ said:
...sed(1) can handle that...
Click to expand...
Something that has been puzzling for a long time and I've been too embarrassed to ask is, what is the significance of the 1 in sed(1)? I've seen things like that for years, and sometimes it's an 8 etc. :q
 
There are multiple sections for man pages that apply to different categories. For example, section 4 is generally device drivers. See the "intro" pages: intro(1), intro(2), intro(3), all the way up to intro(9).

So sed(1) is the man page for sed from the general commands, tools and utilities section.

There usually isn't a lot of overlap in names, so the section number is often left out:
% man sed

But consider re. Is that "re" meaning Perl regular expressions, or "re" the Realtek device driver?
% man 3 re
% man 4 re
 
wblock@ said:
If you know for sure that there will be only spaces and not other whitespace like a tab in front of the numbers, sed(1) can handle that:
% echo "Volume.Capacity_datastore1: 122406567936" | sed -e 's/^.*:\( \)*//'

If there could be tabs, Perl's regex handling is far, far more powerful than sed(1)'s 1972 view of the world:
% echo "Volume.Capacity_datastore1: 122406567936" | perl -pe 's/^.*:(\s)*//'

\s means whitespace, including tabs.
Click to expand...

tr(1) can also be used in conjunct with cut(1) as so:
% echo "Volume.Capacity_datastore1: 122406567936" | cut -d':' -f2 | tr -d ' '

If you need whitespace and tabs tr can do both as well by adding a space before or after \t in this example it's prefixed:
% echo "Volume.Capacity_datastore1:\t\t\t\t 122406567936" | cut -d':' -f2 | tr -d ' \t'
POSIX way has [:blank:] which is both whitespace and tabs which would be % tr -d '[:blank:]'

tr(1) is pretty cool. It's actually one of the older programs that dates back to the multics project.

IT_Architect said:
Both replies were a HUGE help. They both helped me get a handle on working with streams, sed and cut. One good example is worth a 1000 man pages. LOL! I couldn't use them only because they didn't trim the leading spaces in front of 122406567936, but they helped me in my next hurdle tremendously.

Thanks!
Click to expand...

Don't forget to hit the thanks button once in awhile. :)
 
IT_Architect said:
Thanks button? Where?
Click to expand...

It's a button with the thumbs up icon on the lower right side of the post

Image looks like this:
post_thanks.gif
 
You must log in or register to reply here.
Back
Top