All

I have a very large file (aproximately 150,000) as shown below separated by pipe "|". I need to replace data in 2, 16, 17, 23 fields that are of time stamp format. My goal is to look in those fields and it ends with "000000|" then replace it with "000|". In other words, make it as 6 digit micro second value. Some of the records may not end with "000000|". In that case, I shouldn't replace that value. I am enclosing couple of sample records.

Record 1)

154842|2007-11-15 09:33:28.540000000|126074|||02|ALLEN|
DEBORAH|G||2006-06-19 00:00:00|U78BA5|||T|2007-11-15 09:33:28.540000000|2007-11-14 00:00:00|O|O|O|O| |19
00-01-01 00:00:00|2|N/A|DEALLEN

Record 2)

72421|2005-10-07 11:24:00|114307| N/A|Z125A|02|HOLLIS|MENDEZ|D||2005-10-08 00:00:00|U20AWX||64602|A|2005-10-08 00:00:00|2005-11-16 00:00:00|A|A|A|A| |2005-10-0
8 00:00:00|1|Z125|MEHOLLIS

Thanks a lot,
Sekhar

Try...

Ygor:

Thanks for response. It did work. Would you please tell me how it all came together? Especially at \....000\)000/\1. Please let me know this would give us in what we need. Thanks ahead for the help. It really helped me.

sed 's/\(|....-..-.. ..:..:..\....000\)000/\1/g' file1