I have a basic Unix knowledge. I need a script to check and edit a big Invoice file(EDI Message). First the script shall make a small ajustment to the SDT segment.We need to add an ‘A’ at the end of the SDT segment for all flows which don’t denote POP shipments (POP shipments easy to recognize by despatch number which is the RAD number in the EDI message : RAD POP090107). The begining of the file look like follow: //ODETT//INVGM UNB 090114 MID 00001918 090114 SDT012837259 BDT01315000737 PAI EUR 090202 RAD 37062241 081223 CSG01184 CSG02 SDT012837259 to be changed into SDT012837259A when The "RAD" number without "POP" at the begining. Second this big message shall be splittet in two files at a special position. Please let me know if you need more information. Thnaks in advance

I can not give you a complete solution because you don't tell what starts the invoice and what ends it. You need to post more of your data file. The following sed command adds an "A" to the end of any line (which has optional spaces) that starts with the SDT sed '/^SDT/s/[ ]*$/A/' datafile.txt The second part of your requirement is more difficult. The way I read it is that you don't want to append the "A" if a string starting with "RAD" is 3 lines below the line starting with SDT. In shell, this is difficult to do since you have to read ahead of the file until you hit the "RAD" string or the end of the invoice. Personally, I think it's easier to reverse the file so you read the "RAD" string first which tells you whether or not to add the "A". Of course, when the processing is done, reverse the file. That's the best I can do until you reply.

Hi Nails, the begining of this file look like that: //ODETT//INVGM UNB 090106 MID 00000047 090106 SDT012837259 BDT01315000737 PAI EUR 090106 RAD POP081204 081204 CSG0172261 CSG02 ARD0113268023 PRI 000000001NTP000000000000000 ARD0113268176 PRI 000000001NTP000000000000000 ..... ..... //ODETT//INVGM UNB 090106 MID 00000046 090106 SDT012837259 BDT01315000737 PAI EUR 090106 RAD 000081113 081113 CSG0172261 CSG02 ARD0113268019 PRI 000000001NTP000000000000000 ARD0113268022 PRI 000000001NTP000000000000000 RAD POP081114 081114 CSG0172261 CSG02 ARD0113268186 ..... ..... and so on till the end of the file. If you need the complete file I can provide it to you. Like you see there is alot off SDT and RAD segments. But like I told you we need to change(edit) the SDT number only when the "RAD" string number(3 lines below)do not content the "POP" segment at the begining. You're right it is may be simple if we beginn to read first the RAD string. Any replay is welcome. Thanks in advance. Fouad

I have some questions: 1) I'm assuming this string marks the start of an invoice: //ODETT//INVGM 2) How big is the data file, how many invoices typically? 3) Is this statement correct: DO NOT add an "A" to the end of the "SDT" string if the invoice contains a string starting with "RAD"

to 1) it's true to 2) ~24 invoices(sometimes less and sometimes more) to 3) not exactely: DO NOT add an "A" to the end of the "SDT" string if the number after the "RAD" string (3 lines below) do content the "POP" segment at the begining. Example -RAD POP081204 081204--> DO NOT -RAD 000081113 081113 --> DO

As I said, I reversed the invoice file so I could sample the "RAD" & "RAD POP" lines before processing the "SDT" line. Obviously, this won't work if the order is reversed. I used perl to reverse the file and then reverse it back once processing is complete. If you don't want to use perl, check out this article for other reverse options: http://www.samag.com/documents/s=97... newdatafile file contains the desired output. #!/bin/ksh rm -f intdatafile intdatafile2 perl -we 'print reverse <>' datafile > intdatafile newinv=0 while read line do case $line in \/\/ODETT\/\/INVGM*) newinv=0;; esac case "$line" in RAD*) newinv=1;; esac case "$line" in "RAD POP"*) newinv=0;; esac if [[ $newinv -eq 1 ]] then case "$line" in SDT*) newline="$line"A;; *) newline="$line";; esac else newline="$line" fi echo "$newline" done < intdatafile > intdatafile2 # reverse back perl -we 'print reverse <>' intdatafile2 > newdatafile

Hi Nails, I used "awk" to to reverse the file and then reverse it back once processing is complete and it is working. But he result was not correct because this scripts added an "A" to the end of every "SDT" string aside from the "RAD" segment number type(with POP or without). Have you an idea what is going wrong? Many thanks