I need to extract the character before the last "|" in the following lines, which are 'N' and 'U'. The last "|" shouldn't be extracted. Also the no.s of "|" may vary in a line, but I need only the character before the last one.

TT-CAP-N241485-0005241485|ADMINISTRATOR|INC_ORGANISATION_NAME|SUPERIOR_TITLE_1|REV|FIRST_NAME_1|MIDDLE_NAME_1|SURNA ME_FIELD|INC ADDRESSS 1|INC ADDRESSS 2|INC ADDRESSS 3|INC ADDRESSS 4|INC ADDRESSS 5|INC ADDRESSS 6|IN123EC|MALTA|Y||M||Y|LIFEASSURED 2|Y|N||||||||||||||||||||||||Y||Y||||||||||||PRU|A||289832|||FCBGFBF|Y|Y|1|DC|1|1|3|Y|O||20300601||| CAP|5649369|VVUMU7SMT0|HCBCBFDJF|N|Y|N|Y|Y|N|Y|Y|N|3||6|C|20051205|555555555|||||20070604|4|15||C|0| 9|4527|0||20071109|4|C|982062|23569837|36657407|9096||20071205|1|N|

TT-CAP-5241492-0005241492|ADMINISTRATOR1|INC_ORGANISATION_NAME|SUPERIOR_TITLE_1|MS|FIRST_NAME_1|MIDDLE_NAME_1|SURNA ME_FIELD|INC ADDRESSS 1|INC ADDRESSS 2|INC ADDRESSS 3|INC ADDRESSS 4|INC ADDRESSS 5|INC ADDRESSS 6|IN123EC||N||F||N|LIFE ASSURED 1|Y|N||||||||||||||||||||||||Y||Y||||||||||||PRU|BFF||200006|||EB|Y|Y|3|DC|2|1|3|Y|O||20301207|||CAP |5649376|ZLBFQDGVGIOIJDHKNDK|F|N|Y|Y|N|Y|N|N|Y|N|4||7|C|20051205|555555555|||||20070604|2|1600770||B |281215|66|124002|923545||20060405|1|C|1446100|92811|64148521|5||20071205|1|U|

I'm not able to achieve it by cut command. How can it be done ?

you can use awk, set the field separator FS to "|", then use the NF inbuilt variable to get the last field. for the last second field, use NF-1. to get that value, prepend dollar sign, eg $(NF-1)

As the pipe symbol is the last char in the line it is matched by "|$". From there it is easy to construct the following regexp:

sed 's/.*\([NU]\)|$/\1/'

i hope this helps

bakunin

A more general solution using sed is

You are right of course. I wouldn't have limited it if the thread-starter wouldn't have explicitly stated that the character could only be "N" or "U".

Perhaps it is a matter of programming philosophy (which is not the topic of this thread - at least not directly), i would prefer my solution over the general one followed by a line catching all the lines not matched by my solution to point them out as suspicious - lines which "shouldn't exist" at all.

The point is: if a specification says that the specific character can only be "N" or "U", then i would like to be informed if a line with neither "N" or "U" on it comes up. Maybe the specification has changed, maybe my script is faulty, maybe some pre-processing steps producing the lines are faulty - whatever. At least there is *something* that might have to be dealt with.

With all due respect
bakunin

Thanks for your valuable suggestions.