topic Re: find the count in Operating System - HP-UX

find the count

Rahul_13 — Mon, 27 Sep 2004 14:14:41 GMT

Hi,

I have parsed a number of files for the value on a particular field and stored it in a file. Now I want to know the count of each pattern that is stored in the file. For example, the entries in the file has

abc
abc
abc
def
xyz
xyz
123
123
...

and so on. I want to get the count of the occurance of each values. Say, abc - 3 times, 123 - 5 times etc.

Can anyone please help me

Thanks,
Rahul

Re: find the count

Sundar_7 — Mon, 27 Sep 2004 14:18:06 GMT

grep with -c option will tell you the count

# grep -c "abc" inputfile

Re: find the count

Geoff Wild — Mon, 27 Sep 2004 14:28:16 GMT

Bingo! - Sundar has it.

Another way , pipe to wc

grep "abc" somefile |wc

Rgds...Geoff

Re: find the count

Prashant Zanwar_4 — Mon, 27 Sep 2004 14:39:07 GMT

grep pattern | wc -l

this is easiest way to acheive this.
And if you know pattern, you can do some more things of your linking, like

count=1
cat filename | while read line
echo $line
if [ $line = $pattern ]
((count=count+1))
done
something like above you may like to do.
Hope the above helps
Prashant

Re: find the count

john korterman — Mon, 27 Sep 2004 14:42:23 GMT

Hi,
if the lines are adjacant as in your example, you can also try the uniq command, e.g.
# uniq -c
3 abc
1 def
2 xyz
2 123

regards,
John K.

Re: find the count

Mike Stroyan — Mon, 27 Sep 2004 15:27:49 GMT

The general case solution is
sort file | uniq -c
to get a count for each unique line in file.

You can push that through one more sort to order the patterns by frequency.
sort file | uniq -c | sort -bn

Re: find the count

Rahul_13 — Mon, 27 Sep 2004 15:27:52 GMT

Hi,

The values can be anything and it is not known to me so that I could serach for the pattern and do a count of that.

What I need to do is find the count of the occurance of each values which are present in the file. The file also has few blank lines.

Thanks,
Rahul

Re: find the count

Michael Schulte zur Sur — Mon, 27 Sep 2004 15:36:05 GMT

Hi,

what about
grep -v "^ *$" file | sort | uniq -c
?

greetings,

Michael

Re: find the count

Rahul_13 — Mon, 27 Sep 2004 15:42:35 GMT

Hi,

I tried

sed '/^*$/d' < file|sort|uniq -c > newfile

and it worked.

Thanks to everyone.

-Anand.

Re: find the count

Michael Schulte zur Sur — Mon, 27 Sep 2004 15:46:29 GMT

Rahul,

no problems and thanks for taking the time to award points!

Michael

Re: find the count

Prashant Zanwar_4 — Mon, 27 Sep 2004 15:54:16 GMT

YOu can do

filename=yourfilename
cat $filename | sort | uniq >> /tmp/strings.uniq

cat /tmp/strings.uniq | while read line
do
grep -c $line $filename
done

for taking into accoung the blank lines, simply do

grep -c ^$ $filename

Hope above helps

Prashant

Re: find the count

Muthukumar_5 — Mon, 27 Sep 2004 23:26:09 GMT

Remove blank lines by identfying pattern ^$ and do sort and uniq will give results there.

By simply,

sed 's/^$//g' | sort -nb | uniq -c

-nb used to remove leading blanks and compare them there.