[SOLVED] How do I list 'WORDS' containing a specific character? awk,sed,grep?

rohit.dhaval1 · 08-16-2012, 05:14 PM

I have a large file containing thousands of lines, I want to list only strings containig "::".

File example:
Data::Compare 1.22 (1.2101 < 1.22)::
Data:

umper::Concise 2.020 (1.100 < 2.020)
Hash::Merge 0.12 (0.10 < 0.12)::
Path::Class 0.18 (0.16 < 0.18)::
Service-now INC10056109 Install Perl Sort::Topological module Primary 0.02 5.8.8, 5.14.0 TESTING ::
String::Escape 0::
String::Escape 0::
Data::Compare 0::
----------------------
Expected output:
Data::Compare
Data:

umper::Concise
Hash::Merge
Path::Class
Sort::Topological

How can I achieve it?

sycamorex · 08-16-2012, 05:17 PM

What about the line starting with Service? It does contain :: Or the one starting with Path?
What's wrong with it?
You need to be more specific

rohit.dhaval1 · 08-16-2012, 05:25 PM

Quote:

Originally Posted by sycamorex

What about the line starting with Service? It does contain :: Or the one starting with Path?
What's wrong with it?
You need to be more specific

No matter how the line starts or ends, all I need is list of words containing '::' in it.

sycamorex · 08-16-2012, 05:31 PM

Try:

Code:

awk '$1 ~/::/ {print $1}' file

sycamorex · 08-16-2012, 05:40 PM

My solution above may work or may not. It's not exactly what I think you're after.

Is that what you want?

Code:

Data::Compare
Data:umper::Concise
Hash::Merge
Path::Class
Sort::Topological
String::Escape
String::Escape
Data::Compare

rohit.dhaval1 · 08-16-2012, 05:45 PM

Many thanks.. That is exactly what I wanted.

sycamorex · 08-16-2012, 05:46 PM

What about that:

Code:

awk '{for (i=1;i<=NF;i++) if ($i ~/[a-z]::/) print $i}' file

sycamorex · 08-16-2012, 05:49 PM

Quote:

Originally Posted by rohit.dhaval1

Many thanks.. That is exactly what I wanted.

Please note that the solution from post #4 will NOT give you the output from post #5
Try the solution from post #7

grail · 08-16-2012, 08:31 PM

How about a little gem (pun):

Code:

ruby -ne 'puts $_.scan(/(\w+(::\w+)+)/)[0][0]' file

rohit.dhaval1 · 08-21-2012, 08:00 PM

thank you all

David the H. · 08-22-2012, 09:08 AM

Code:

$ grep -Eo '\b[^ ]+::[^ ]+\b' inputfile.txt
Data::Compare
Data:umper::Concise
Hash::Merge
Path::Class
Sort::Topological
String::Escape
String::Escape
Data::Compare

And please use ***[code][/code] tags*** around your code and data, to preserve formatting and to improve readability. Please do not use quote tags, bolding, colors, or other fancy formatting.

If you go into the advanced editing box, there's also an option allowing you to turn off the smiley faces.