Grep to match unknown pattern

Shell Programming and Scripting
1

Hi there

I would like to search a file for a certain pattern, but i don't know the exact pattern i need to search for.

What i do know is that i need to search for the pattern that exists the most in a certain file.

so if a file looksike this:

summer, summer, winter, spring, summer summer.

i would need a grep command to display summer, while summer is the pattern that exists the most.

regards chilly

2

I think instead you want something to count occurences of unique words, That would be awk or perl, not grep.

3

OK, I'll bite :rolleyes::

cat infile:

Hi there
I would like to search a file for a certain pattern, but i don't know the exact pattern i need to search for.
What i do know is that i need to search for the pattern that exists the most in a certain file.
so if a file looksike this:
summer, summer, winter, spring, summer summer.
i would need a grep command to display summer, while summer is the pattern that exists the most.
regards chilly

$> grep $(grep -o '\b[^[:space:]]*\b' infile |awk '{A[$1]++}END{for (i in A) print A,i}'|sort -n|tail -1|cut -d ' ' -f2) infile
summer, summer, winter, spring, summer summer.
i would need a grep command to display summer, while summer is the pattern that exists the most.

or do you mean:

$> grep -o '\b[^[:space:]]*\b' infile |awk '{A[$1]++}END{for (i in A) print A,i}'|sort -n|tail -1|cut -d ' ' -f2
summer
4

Thx for the replies.

Since in a newbie to Linux i hardly understand scrutinzers answer, although it's a very impressive command line :confused::wink:

@Tony: have you got an example of how i would locatie the words that exist the most?

5

The word that exist the most..

quick version could be this.. in perl.

while(<>)  {
    @arr = split /\b/, $_;
    foreach (@arr)  {
        $hash{$_}++ if not ( /\s+/ || /,/ );
    }
}

@arr = sort { $hash{$b} <=> $hash{$a} } keys %hash;
print $arr[0];