example of data file

BEY
BEY
BEY
BEY
BEY
BEY
00219 ELC03/306739 (i want to remove the patterns like this how to do)

00219 IC03/502508

00219 (TBI)

00219 PNL04/601784

TEB
TEB
TEB
TEB
TEB
TEB

If that string only occurs once, you could fire up an editor and remove the line.

Both
grep -v '00219 ELC03/306739' <insert filename here>
and
sed '/00219 ELC03\/306739/d' <insert filename here>
(notice the backslash) will output the contents of the data file with all lines which contain the string "00219 ELC03/306739" removed. You'll need to do more processing if you want the string removed from the original file.

Hi My data file islike this

7777777711|SIN|ESC|SIN||2003-09-08 11:00:00|D|FOCMATHUB|FOCMATHUB|DHL/SIN HUB OF
FICE|819458|TEST||TEST DATA|TEST DATA|TEST DATA|TEST DATA|TEST DATA|SINGAPORE|EX
PRESS DOCUMENT|DOX||TEST|75 AIRPORT CARGO ROAD|SINGAPORE|||||||||TEST DATA||||SI
NGAPORE||SG|
I used
gzcat Jan05.gz | awk 'BEGIN { FS = "|" } ; { print $2 }' > cnt to extract the second column

I need only

SIN
BEY
SIN
BEY
BEY
BEY

But my file cnt contains

BEY
BEY
BEY
BEY
00219 ELC03/306739
00219 IC03/502508
00219 (TBI)
BEY
BEY
BEY

I want to remove the 00219 ELC03/306739
00219 IC03/502508
00219 (TBI) contents how to do it? i WANT TO DO IT
AN GENERIC MANNER AND NOT PARTICULARY THIS TEXT ANY NUMBER OR CHARS MORE THAT 3 I WANT TO REMOVE IT HOW TO DO?

The tool that comes to my mind with tasks like this is 'Awk.' I don't know exactly what the acronym stands-for but it's really useful for working with complex inputs a line at a time. For example, it's handy when all you've got is the output of a COBOL report (saved to a disk file) and you're supposed to extract some data from it.

Another tool might be 'grep' (global regular expression pooh-bah).

A surprisingly useful programming language for working with files like this is also, Python. If you can use Perl then be my guest. ;-)

I used awk and extracted the data file and i want to remove the numeric
contents that can done with sed . I dont know how to do it in generic
maaner. please tell me.

awk - The name awk comes from the initials of its designers: Alfred V. Aho, Peter J. Weinberger, and Brian W. Kernighan AWK

ok----

regards
s.kalyan

The pattern match selects fields of length exactly three characters:

awk 'BEGIN{FS="|";} $2 ~ /^...$/ { print $2; }'