Shell Programming and Scripting

Hi,

I have a file (log data) with the sample code as follows

/apps/opt | go_gen.ksh | #START | jan 28 06:15 |/home/dump
/apps/opt | go_gen.ksh | #END | jan 28 06:30 |/home/dump

/apps/opt | pend_instl.ksh | #START | jan 28 09:50 |/home/dump

/apps/opt | pend_instl.ksh | #END | jan 28 09:55 |/home/dump

/apps/opt | go_gen.ksh | #START | jan 28 11:30 |/home/dump

/apps/opt | go_gen.ksh | #END | jan 28 11:35 |/home/dump

/apps/opt | create_dump.ksh| #START | jan 28 12:51 |/home/dump

/apps/opt | create_dump.ksh| #END | jan 28 12:52|/home/dump

/apps/opt | go_gen.ksh | #START | jan 28 11:30 |/home/dump

/apps/opt | go_gen.ksh | #START | jan 28 15:30 |/home/dump
/apps/opt | go_gen.ksh | #END | jan 28 15:35 |/home/dump
/apps/opt | installed.ksh| #START | jan 28 15:56 |/home/dump

/apps/opt | installed.ksh| #END | jan 28 15:59 |/home/dump
.
.
.
(100s of logs of same format)
I need to write a shell program to create a file which contains
(1st field (file path),2nd field (file name),4th field(start time),end time (in line having #END) and 5th field(log path)..for each filename

for ex: take line1 and line2 in the sample data i mentioned..they corresponds to on file name,first line giving star time and second line giving end time.

The Problem is
1)for certain files like go_gen.ksh (highlighted as bold),there are more tha one entry in the inputlog file.I need to generate corresponding start and end times(which i am unable to get while writing in a while loop) in the new file.
2) for certain filenames like "/apps/opt | go_gen.ksh | #START | jan 28 11:30 |/home/dump" there is no corresponding ENDTIME.In such case i need to put a '-' in the endtime field in newfile.


Need to write a shell program.Please can u give me the outline and rough code.Since i am new to shell programming,i am unaware of most of the functions/commands.(even awk/sed is permisable)

Reply me as soon as possible.


Thanks in anticipation.

Could you post an example the expected output?

The output should like this..(for the data mention above)

/apps/opt |go_gen.ksh |jan 28 06:15 |jan 28 06:15 |/home/dump
/apps/opt |pend_instl.ksh |jan 28 09:50 |jan 28 09:55 |/home/dump
/apps/opt |go_gen.ksh | jan 28 11:30 |jan 28 11:35 |/home/dump
/apps/opt | create_dump.ksh|jan 28 12:51 |jan 28 12:52|/home/dump
/apps/opt | go_gen.ksh |jan 28 11:30 |- |/home/dump
apps/opt | installed.ksh| jan 28 15:56|jan 28 15:59 |/home/dump

Given your example input you may try something like this:

[use gawk, nawk or /usr/xpg4/bin/awk on Solaris]

Code:

awk -F\| 'END { 
  print s, t[1], t[2] ? t[2] : " - ", l 
  }
/#START/ && s { 
  print s, t[1], t[2] ? t[2] : " - ", l 
  split(FS, t); c = 0
  }  
NF { s = $1 FS $2; l = $5; t[++c] = $4 }
' OFS=\| infile