Parsing a variable length record

Barb · #1 (**permalink**) 09-28-2004

I need to pick a field out of a variable record - the field is always found 4 fields after a certain text string, but it can be on any line of the record and in any position across the record on a line. I have had no luck through any of the Unix editors being able to cut a field that isn't always at the same spot in a record. Does anyone have any ideas - or do you know how I can replace a character string in a file with a hard return?

Perderabo · #2 (**permalink**) 09-28-2004

This sounds either very hard or downright impossible. We need a lot more info. Is the data file a collection of lines, each ending with a new line character? How do you tell where a field ends? How do you tell where where a record ends? How about some sample data?

Barb · #3 (**permalink**) 09-28-2004

Here is a sample of the file - the '~ST' appears at the end of every record. What I need to pull out is the 5th field from the ~ST (fields are seperated by '^') based on the value of 855 or 856 in the 1st field. I know how to use awk to search for a character string and return values in fields on the line. But I have never had to start any of the editors with lines over multiple rows and fields all over the place. And a record can be any length. I feel like I've entered the twilight zone. Thanks for any help you can give me on this.

Barb · #4 (**permalink**) 09-28-2004

Here is the actual sample - whoops

Code:

   ~REF^BO^C61102^BIN LOCATION~PO1^29^6.00^PK^9.50^^VC^435916^IN^916~PID^F^^^^PAD, DRI FLOW (12/PK 6PK/CS)~ACK^AC^6.00^PK~REF^
AP^C400=PK   ^          ~REF^BO^C61102^BIN LOCATION~CTT^29^334.00~SE^139^0001~GE^1^5644~IEA^1^000005701~@@@@@@@@@@@@@@@@@@@@@@@@
ISA^00^D011-56018^00^VAX-VMSSYS^01^GENERAL MEDICAL^ZZ^011            ^040506^1702^U^00200^000005702^0^P^|~GS^PR^GM^56018^040506^
1702^5645^X^003020~ST^855^0001~BAK^06^AC^440597^040506^^^^^040506~REF^IV^723065~DTM^055^040506^1703~PER^BD^JANET ARROYO~N1^ST^^9
1^573~N1^BT^0100^91^56018~PO1^1^12.00^BX^7.39^^VC^213458^IN^15074~PID^F^^^^ELECTRODE, EKG INSTATRACE 5'S (50/BX)~ACK^AC^12.00^BX
~REF^AP^C72=BX    ^          ~REF^BO^C61802^BIN LOCATION~CTT^1^12.00~SE^14^0001~ST^855^0002~BAK^06^AC^440598^040506^^^^^040506~R
EF^IV^723066~DTM^055^040506^1703~PER^BD^SAE JACKSON~N1^ST^^91^657~N1^BT^0100^91^56018~PO1^1^12.00^EA^0.28^^VC^278487^IN^1207~PID
^F^^^^BATTERY, ALKALINE "AAA" (24/BX 3BX/CS)~ACK^AC^12.00^EA~REF^AP^C60=BX    ^          ~REF^BO^C61202^BIN LOCATION~PO1^2^24.00
^EA^0.47^^VC^278484^IN^127~PID^F^^^^BATTERY, ALKAL "C" (SPEC) (12/BX 6BX/CS)~ACK^AC^24.00^EA~REF^AP^C100=BX   ^          ~REF^BO
^C61802^BIN LOCATION~PO1^3^48.00^EA^0.25^^VC^278485^IN^130~PID^F^^^^BATTERY, ALKAL "AA" (24/BX 6BX/CS)~ACK^AC^48.00^EA~REF^AP^C3
02=BX   ^          ~REF^BO^C61902^BIN LOCATION~PO1^4^6.00^EA^0.58^^VC^186661^IN^780~PID^F^^^^SOD CHL, IVSOL 0.9% 500ML (24/CS)~A
CK^AC^6.00^EA~REF^AP^C305=CS   ^          ~REF^BO^C61102^BIN LOCATION~CTT^4^90.00~SE^29^0002~GE^2^5645~IEA^1^000005702~@@@@@@@@@
ISA^00^D011-56018^00^VAX-VMSSYS^01^GENERAL MEDICAL^ZZ^011            ^040506^1801^U^00200^000005703^0^P^|~GS^PR^GM^56018^040506^
1801^5646^X^003020~ST^855^0001~BAK^06^AC^440601^040506^^^^^040506~REF^IV^723079~DTM^055^040506^1803~PER^BD^SCOTT FITZGERALD~N1^S
T^^91^609~N1^BT^0100^91^56018~PO1^1^15.00^EA^6.08^^VC^405660^IN^14789~PID^F^^^^C

Perderabo · #5 (**permalink**) 09-28-2004

Whatta mess...

Maybe this will get it...

sed -n '/\~ST\^/s/.*\~ST\^//p' < inputfile | awk -v FS=^ '$1=855 { print $5 }'

Barb · #6 (**permalink**) 09-28-2004

Perderabo, you are wonderful - Thank you very much it worked perfectly.

rhfrommn · #7 (**permalink**) 09-28-2004

Wow.

That was pretty amazing Perderabo, being able to figure out how to extract the data from that mess - and getting it right on the first try.

More UNIX and Linux Forum Topics You Might Find Helpful
Thread	Thread Starter	Forum	Replies	Last Post
What the command to find out the record length of a fixed length file?	tranq01	UNIX for Dummies Questions & Answers	9	12-04-2008 04:04 PM
Pivot variable record length file and change delimiter	thomasr	Shell Programming and Scripting	3	10-08-2007 08:40 PM
fixed record length	george_	Shell Programming and Scripting	16	03-28-2006 05:41 AM
creating a fixed length output from a variable length input	r1500	Shell Programming and Scripting	2	12-03-2003 01:09 PM
Parsing data and retaining the full length of variable	app4dxh	Shell Programming and Scripting	3	11-22-2002 12:04 PM

Thread Tools	Search this Thread
Show Printable Version Email this Page	Search this Thread: Advanced Search
Display Modes	Rate This Thread
Switch to Linear Mode Hybrid Mode Switch to Threaded Mode	Rate This Thread: