Counting number of occurance in awk program

sarajevo · 10-31-2006, 07:29 AM

Hi all, I am hoping that someone will have time to read about my problem and eventually write down some answer

I have some huge file and I want to check it on some values so I wrote script bellow
###################################################################
cat 942prm_$d | tr ' | ' ' ' > prm942_$d.txt # to translate | into empty space
d=`TZ=MET+24 date +%Y%m%d` # date

# And awk rule...

awk ' { if ( ($10==50005) && ($22==128) && ( $12~/^61/) || ($12~/^62/) && length($12) <= 8) print $10; print $12 }' prm942_$d.txt > mobilna.txt

awk ' { if ( ($10==50006) && ($22==128) && ( $12~/^61/) || ($12~/^62/) && length($12) <= 8) print $10; print $12 }' prm942_$d.txt > mobilna.txt

awk ' { if ( ($10==50007) && ($22==128) && ( $12~/^61/) || ($12~/^62/) && length($12) <= 8) print $10; print $12 }' prm942_$d.txt > mobilna.txt
#######################################################################

if colummn $10 is 50005 and $22 is 128 and $12 begins with 61 or 62 and length of that line is not smaller that 8 I need to print colummn $10 and $12 and I made it with above awk command. Where is column $1 represent called number and column $12 is number who called number in $1. $12------>$10 ( direction of call )
Now I need just small hint ( help ) how count how many times each case occur. I need output like this,

50005---------------4 ( means: the awk rule was ok 4 times for number 50005 )
50006---------------5 ( as above, just five times for number 50006)
50007---------------6 ( as above, just six times for number 50007)

in file mobilna.txt.
In other words how many times some customer which phone number begins with 61 or 62 called number 50005—first awk command, number 50006---second awk command, 50007 --- third awk command and so on.
What I want to say, if rules in awk command

awk ' { if ( ($10==50005) && ($22==128) && ( $12~/^61/) || ($12~/^62/) && length($12) <= 8) print $10; print $12 }' prm942_$d.txt > mobilna.txt

occurs some number of times for number 50005, I need to get output like above. I need to know how many time occurs that ( that is number of calls on that number ). I probably need some kind of counter but I do not know how implement that.
Every help or hint is welcome.
Thanks

card-suse · 10-31-2006, 08:29 AM

awk has the ability to do arrays. Indices for arrays in awk are actually characters. So you can define an array like:

Code:

++Counter[$10]

This should simultaneously allocate and increment the counter array if it doesn't already exist. I'm not sure if awk initializes the value to zero and I don't have the time right now to check. It it doesn't, then you need to initialize everything in a BEGIN block. You'll have to put your awk code in a file since it will take a few lines to do what you want to do. But here's an example (make a file called "myprog.awk"):

Code:

# This is the BEGIN block that initializes stuff

BEGIN{
Counter[50005]=0;
}

# Here is your conditional statement

if ( ($10==50005) && ($22==128) && ( $12~/^61/) || ($12~/^62/) && length($12) <= 8) 
{
  Counter[50005]+=1
}

# This is the END block that will output the results

END{
for (i in Counter) 
{
  printf("%10d--------%10d",i,Counter[i]
}

You could then run this file with:

Code:

awk -f mycode.awk prm942_$d.txt

That should do it....

card-suse · 11-01-2006, 10:31 AM

Well, did that work for you???????