awk commands/script to extract lines with a specific format

oldfogey · 10-02-2012, 10:42 AM

Hi. I'm trying to extract lines from a file, each of which are of a certain format. I can write awk commands to do parts of this as individual steps but I'd like to link everything together so I don't have to issue a sequence of individual commands.

Here's a sample portion of the input file:

F9JcQEcS21=SQZK6if6mfnA
pc\sender4
Wed12:25:17
abc124
ba1
clp1491
N18471
Tam-port
dd
e.44
flash9
1c-64-94-98-87-11

I'd like to extract all lines from the input file that:

1) only contain alphanumeric characters (all alpha characters a-z being lowercase)
2) the beginning of each line must have a minimum of 2 alpha characters
3) the last character must be a number.

So the properly extracted lines from the sample above would be:

abc124
ba1
clp1491
flash9

Thanks in advance for your suggestions!

rigor · 10-02-2012, 12:16 PM

Based on what you show as desired output, it sounds as if you want to output lines that start with a minimum of two lower case letters, end with digits, and contain only lower case letters and digits. Using your example data as input, this command gets the desired output:

Code:

gawk --posix '/^([a-z]{2,})([0-9]+)$/ { print $0 }'

cortman · 10-02-2012, 12:26 PM

You can also do this with grep-

Code:

egrep "^([a-z]){2}[a-z]*[0-9]*[0-9]$"

then direct the output to a file or xargs.

oldfogey · 10-02-2012, 12:34 PM

Fantastic! You guys are tremendous!

Thank you.

SecretCode · 10-02-2012, 01:39 PM

Quote:

Originally Posted by kakaka

Based on what you show as desired output, it sounds as if you want to output lines that start with a minimum of two lower case letters, end with digits, and contain only lower case letters and digits. Using your example data as input, this command gets the desired output:

Code:

gawk --posix '/^([a-z]{2,})([0-9]+)$/ { print $0 }'

This would produce the example output, but the description given would also allow digits followed by letters - lines like abc123def99

If the description's correct and the examples were missing such a case, you would need

Code:

gawk --posix '/^[a-z]{2}[a-z0-9]*[0-9]+$/ { print $0 }'

oldfogey · 10-02-2012, 01:50 PM

You're correct. A string such as abc123def99 would not be desired in the output.

The output must be (minimum 2) alphas followed by (at least 1) numeric.

Thanks for the catch!