count pattern occurrences in a line

ymho · #1 (**permalink**) 01-21-09, 11:26

pls advise how to count pattern occurrences in a line.

#:#1#:#2#:#3#:#4
=> no of #:# : 4

i try
echo "#:#1#:#2#:#3#:#4" |sed 's/#:#/\n/g ' |grep -c ...

but doesnt work in solaris

thx

mike_bike_kite · #2 (**permalink**) 01-21-09, 11:43

This should work:

Code:

echo "#:#1#:#2#:#3#:#4" | awk '{ split($0,a,"#:#"); print length(a) - 1 }'

ymho · #3 (**permalink**) 01-21-09, 11:51

good good good ... thx mike_bike_kite again ... but u seem love to use awk a lot ...

mike_bike_kite · #4 (**permalink**) 01-21-09, 12:25

whatever works

vgersh99 · #5 (**permalink**) 01-21-09, 19:48

Quote:

Originally Posted by ymho

pls advise how to count pattern occurrences in a line.

#:#1#:#2#:#3#:#4
=> no of #:# : 4

Actually there're 5 (five) fields in this record: '', '1', '2', '3', '4'

but if don't want to count the first 'empty' field, you can always substract '1':

Code:

 echo "#:#1#:#2#:#3#:#4" | awk -F'#:#' '{ print NF-1}'

mike_bike_kite · #6 (**permalink**) 01-22-09, 06:37

Yep - that's a better way of doing it

Internet Services

Web Development

Digital Marketing

Consumer Tech