Question

I have a file with different lines, among which I have some lines like

173.194.034.006.00080-138.096.201.072.49934

the pattern is 3 numbers and then a dot and then 3 numbers and then a dot, etc.

I want to use awk, grep, or sed for this purpose. How do I express this regular expression?

Answer 1

Assuming you want to get lines with 1 series like 123. exists, do

 grep '[0-9][0-9][0-9]\.' file > numbersFile

If you want 2 series like 123.345., then do

 grep '[0-9][0-9][0-9]\.[0-9][0-9][0-9]\.' file > numbersFile

etc, etc.

Each [0-9] means match only one occurance of characters in the range between 0-9 (0,1,2,3,4,5,6,7,8,9).

Because the ‘.’ char has a special meaning in a normal grep regexp, you nave to escape it like \. to indicate “Just match the ‘.’ char (only!) 😉

There are fancy extensions to grep that allow you to specify the pattern once, and include a qualifier like {3} or sometimes \{3\} (to indicate 3 repetitions). But this extension isn’t portable to older Unix like Solaris, AIX, and others.

Here’s a simple test to see if your system supports qualifiers. (Super Grep-heads are welcome to correct my terminology :-).

   echo "173.194.034.006.00080-138.096.201.072.49934" | grep '[0-9]\{10\}\.'
   echo "173.194.034.006.00080-138.096.201.072.49934" | grep '[0-9]\{2\}\.'

The first test should fail, the 2nd will succeed if your grep supports qualifiers.

It doesn’t hurt to learn the long-hand solution (as above), and you can be sure this will work with any grep.

IHTH.