Removing duplicates in a sorted file by field.

Shell Programming and Scripting
1

I have data like this:
It's sorted by the 2nd field (TID).
envoy,90000000000000634600010001,04/11/2008,23:19:27,RB00266,0015,DETAIL,ERROR,
envoy,90000000000000634600010001,04/12/2008,04:23:45,RB00266,0015,DETAIL,ERROR,
envoy,90000000000000634600010001,04/12/2008,23:14:25,RB00266,0015,DETAIL,ERROR,
envoy,90000000000000634600010001,04/13/2008,04:23:39,RB00266,0015,DETAIL,ERROR,
envoy,90000000000000634600010001,04/13/2008,22:41:58,RB00266,0015,DETAIL,ERROR,
envoy,90000000000000634600010001,04/13/2008,22:42:44,RB00266,0015,DETAIL,ERROR,
envoy,90000000000000634600010001,04/13/2008,22:49:43,RB00266,0015,DETAIL,ERROR,
envoy,90000000000000634600010001,04/13/2008,22:50:45,RB00266,0015,DETAIL,ERROR,
envoy,90000000000000634600010001,04/13/2008,22:53:23,RB00266,0015,DETAIL,ERROR,
envoy,90000000000000634600010001,04/14/2008,12:38:40,RB00266,0015,DETAIL,ERROR,
envoy,90000000000000634600010001,04/14/2008,12:52:22,RB00266,0015,DETAIL,ERROR,
envoy,90000000000000693200010001,04/17/2008,09:07:09,RB00060,0009,ENVOY,ERROR,26
envoy,90000000000000693200010001,04/18/2008,10:27:13,RB00083,0009,ENVOY,ERROR,26
envoy,90000000000000693200010001,04/18/2008,11:36:27,RB00084,0009,ENVOY,ERROR,26
envoy,90000000000001034800010001,04/01/2008,23:59:15,RB00294,0030,ENVOY,ERROR,57
envoy,90000000000001034800010001,04/02/2008,23:59:12,RB00295,0030,ENVOY,ERROR,57
envoy,90000000000001034800010001,04/03/2008,23:59:11,RB00296,0030,ENVOY,ERROR,57
envoy,90000000000001034800010001,04/04/2008,23:59:08,RB00297,0030,ENVOY,ERROR,57
envoy,90000000000001034800010001,04/05/2008,23:59:04,RB00297,0030,ENVOY,ERROR,57
envoy,90000000000001034800010001,04/06/2008,22:59:06,RB00297,0030,ENVOY,ERROR,57

I want to do the following:
Check the second field to see if the TID is the same as the previous line. If the TID has been seen before then check the 7th field to see if that is the same as the previous line. If both are the same, I want to remove the line and increment a counter.

My ideal output would look something like this.
11,envoy,90000000000000634600010001,04/11/2008,23:19:27,RB00266,0015,DETAIL,ERROR,
3,envoy,90000000000000693200010001,04/17/2008,09:07:09,RB00060,0009,ENVOY,ERROR,26

etc.

I figure I actually need to do an awk script rather than a 1 liner. The other option is for the last 3 fields to be one field and compare by the TID field and the error field, then split them into 3 on the output.

Any thoughts? I've looked at other stuff removing dups with awk and it's mostly one liners.

I'd love to ask get some explanation of WHY it works so that I can mod it if need be.

2

For the first question you can use the 2nd and the 7th field in an associative array:

awk -F, '
!arr[$2,$7]{arr[$2,$7]=$0;c[$2,$7]++;next}    # If the array is new, preserve the first line and increase the counter
arr[$2,$7]{c[$2,$7]++}                        # else just increase the counter
END{for(i in arr){print c","arr}}       # Print the counter with the first line
' file

Give an example of the expected output for your other option.

Regards