Shell Programming and Scripting

I have a file which contains records in the format of

2006-08-25 12:06:13|ABC|93
2006-08-25 12:45:55|ABC|203
2006-08-25 01:48:19|DEF|156
2006-08-25 01:49:09|ABC|12798
2006-08-25 02:49:59|GHL|4109
2006-08-25 03:50:50|DEF|234

where the format is "arrive time"|"message type"|"processing time"

I want to find out
- number of records in the file
=> 6 for above case
- number of records in the file for each message type
=> ABC = 3
DEF = 2
GHL = 1
- distribution of message arrive time
=> 12 - 13 = 2
13 - 14 = 2
14 - 15 = 1
15 - 14 = 1
- average processing time
=> 2932.2

Can anyone helps me with this? THANK YOU!

--To find the number of line in the file use
awk 'END{print $NR}'
--to get which message repeated how many time i believe you can follow these step
* cut down the 2nd column sepearated using "|"
say awk 'BEGIN{FS=OFA="|"}{print $2}' srcfile.txt > src_temp.txt
* then sort the src_temp.txt to get the uniq message
sort -o src_temp.txt src_temp.txt | uniq
now search the sorce file using src_temp.txt as your pattern file which actually contain the uniqmessage name

*Probably what you can do is

while read msg
do
{
msgcnt =`grep -c $msg src_file.txt`
echo "$msg = $msgcnt"
}done < src_temp.txt

rm -fr src_temp.txt
-----------------------

I am confused of your third requirement

jambesh, what I want to be able to find out how many records were logged between each time range, I have set the range as a hourly interval, I am hoping to use the timestamp in the first part of the record to distinguish which range is fall into then increment that count. Thanks

See my comments re tis message sent in by "jerrad" "Make grep -c display like grep -n? "
It has a quick little script that will out put (Total times Appears):string
ie
3:ABC
2EF
etc etc