UNIX question--please help!!!

[aircasper]

i don't know where else to turn to at this point, so i thought i'd turn to the helpful (and knowledgeable) anandtech community. my wife is taking a unix class (don't know why you need that for a PhD in Sociology) and she needs to figure out what command to use in order to count how many times a specific word appears in a file.

i guess the grep -c command is not what she needs, since that only counts how many lines the word appears in, as opposed to how many times the word appears in the file (i.e., accounting for multiple occurrences of a word within a single line). i've been all over the web with google trying to find an answer, with no luck. any thoughts guys and gals????

thanks in advance!

 

[Talon]

something like this:

grep the_word_being_searched | wc

use grep to search for the specific word, pipe to wc for word count

scroll down to pipes and redirection at

this site. It has a better explanation.

 

[aircasper]

thanks talon! my wife doesn't want me to forward her the answer yet, as she wants to try and figure it out. but i gave her the hint about piping and redirection based on your help. it's kind of funny how i'm the one who's more into computers and such, but my wife (who is pratically anti-computer) is turning into the unix geek (not on her own accord of course).


hmm, it doesn't seemed to have worked. my wife said that was actually one of the first things she tried. apparently, it gives her a word count for the lines that contain the word she's looking for, rather than telling her the number of times the specific word appears.

 

[KK]

Originally posted by: Talon
something like this:

grep the_word_being_searched | wc

use grep to search for the specific word, pipe to wc for word count

scroll down to pipes and redirection at

this site. It has a better explanation.

Will this give you a word count for the lines that have the searched word in them and not of how many time the searched word was found.

KK

 

[Talon]

KK may be correct. So much for that other site I posted a link to.

 

[aircasper]

yeah, unfortunately it didn't work. but thanks for the help talon. any other thoughts?

 

grep -o -i <word> <file> | wc

 

[notfred]

perl -e 'open F, "filename.txt";while(<F> ) {while ($_ =~ /word/g){$i++}}print "$i\n";'

replace "filename.txt" with the name of your file and "word" with the word you're searching for.

 

[Legendary]

Originally posted by: notfred
perl -e 'open F, "filename.txt";while(<F> ) {while ($_ =~ /word/g){$i++}}print "$i\n";'

replace "filename.txt" with the name of your file and "word" with the word you're searching for.

This must be the simplest Perl possible because I actually understand what it means without ever seeing Perl before.
/me hands himself a cookie.

 

[polm]

cat file |fmt -w1 |grep word |wc -l

 

[polm]

Originally posted by: polm
cat file |fmt -w1 |grep word |wc -l

replace file with actual filename
replace word with the word you want a count of

 

[wyvrn]

Not wc -l , (-l counts lines, not words).

grep -o -i <word> <file> | wc -w

what will appear on the screen is the number of times the word was found

 

[polm]

Originally posted by: wyvrn
Not wc -l , (-l counts lines, not words).

grep -o -i <word> <file> | wc -w

what will appear on the screen is the number of times the word was found

that is why I included the | fmt -w1 |

 

[The Dancing Peacock]

Originally posted by: polm

Originally posted by: polm
cat file |fmt -w1 |grep word |wc -l

replace file with actual filename
replace word with the word you want a count of

This didn't work for me.

I have a file called hungup with about 20 lines, each with having the word root in the line, and here was my output

cat hungup |fmt -w1|grep root|wc -l
fmt: bad width: 0
0

 

[Chaotic42]

Originally posted by: The Dancing Peacock
This didn't work for me.

I have a file called hungup with about 20 lines, each with having the word root in the line, and here was my output

cat hungup |fmt -w1|grep root|wc -l
fmt: bad width: 0
0

I made a file called rootfile, put 20 "roots" in it, one on each line, and here's what I got.
/tmp$ cat rootfile |fmt -w1 |grep root |wc -l
20

Do you have the spaces in the right places?

 

[notfred]

Originally posted by: Legendary

Originally posted by: notfred
perl -e 'open F, "filename.txt";while(<F> ) {while ($_ =~ /word/g){$i++}}print "$i\n";'

replace "filename.txt" with the name of your file and "word" with the word you're searching for.

This must be the simplest Perl possible because I actually understand what it means without ever seeing Perl before.
/me hands himself a cookie.

So could you explain to me what it does,then?


It appears that my perl script is the only thing that works so far

 

[polm]

Originally posted by: Chaotic42

Originally posted by: The Dancing Peacock
This didn't work for me.

I have a file called hungup with about 20 lines, each with having the word root in the line, and here was my output

cat hungup |fmt -w1|grep root|wc -l
fmt: bad width: 0
0

I made a file called rootfile, put 20 "roots" in it, one on each line, and here's what I got.
/tmp$ cat rootfile |fmt -w1 |grep root |wc -l
20

Do you have the spaces in the right places?

 

[gopunk]

you probably need it for statistical stuff

 

[Rias]

Originally posted by: notfred
perl -e 'open F, "filename.txt";while(<F> ) {while ($_ =~ /word/g){$i++}}print "$i\n";'

replace "filename.txt" with the name of your file and "word" with the word you're searching for.

cat filename.txt | perl -nle 'while (m/\bword\b/ig) { $i++ } END { print $i }'

or

cat filename.txt | perl -nle '$i++ while m/\bword\b/ig; END { print $i }'

will only catch if the word is "word" not "wordplay" , and it's case insensitive.

 

[aircasper]

wow, a lot of suggestions here. i'll have my wife try them out when we get home (stuck in the library stacks right now). thanks everyone!