Regular expression - filename with multiple periods

Question

Consider the following command line: tfile -a -fn P2324_234.w07 -tc 8811

The regex to parse this: -\w+|\w+\s|\w+\.+\w+\s (see screenshot below)

The problem is when the file name has multiple dots, say: tfile -a -fn P23.24.23.4.w07 -tc 8811

Question: how to ensure the P23.24.23.4.w07 is parsed as one argument (as in P23.24.23.4.w07)?

enter image description here

Have a read [here](http://stackoverflow.com/questions/17043454/using-regexes-how-to-efficiently-match-strings-between-double-quotes-with-embed). It essentially boils down to the same problem. Also, you might want to use an online tester, that supports the flavor you are actually using: http://regexplanet.com — Martin Ender, Jun 12 '13 at 01:24

Casimir et Hippolyte · Accepted Answer · 2013-06-12T01:40:28.717

3

Describe it!

For: P23.24.23.4.w07
use: \w+(?:\.\w+)+

note that for your java version you can use possessive quantifiers and atomic groups:

\\w++(?>\\.\\w++)+

edited Jun 12 '13 at 01:40

answered Jun 12 '13 at 01:24

Casimir et Hippolyte

88,009
5
94
125

But that's an evil regex. See http://stackoverflow.com/questions/12841970/how-can-i-recognize-an-evil-regex. – AJMansfield Jun 12 '13 at 01:25
4

@AJMansfield no it's not. There is no overlap between `\w` and `.` which makes it pretty safe. This is called "unrolling-the-loop" and is one of *the* efficiency techniques. See the link in my comment on the question. – Martin Ender Jun 12 '13 at 01:27
@AJMansfield `(a+)+` can be source of catastrophic backtracking, but (ab+)+ is not that dangerous. +1 m.buettner. – Pshemo Jun 12 '13 at 01:31

score 3 · Answer 2 · answered Jun 12 '13 at 01:29

3

Use a character class, e.g., /-fn [a-z0-9.]+ -tc/i. In English, that means "-fn, followed by one or more of characters between a-z, between 0-9, or a ., followed by -tc." If you want to capture that part, wrap that part in parentheses.

answered Jun 12 '13 at 01:29

thomasd

2,593
1
22
23

You should probably also include underscores, and then this implifies to `[\w.]`. – Martin Ender Jun 12 '13 at 01:30

score 1 · Answer 3 · answered Jun 12 '13 at 01:36

1

I have used this -\w+|\w+\s|\S+.+\w+\s

Instead of 'word', we may use 'not space', You have not specified your extra requirement so I think it is fine.

Regexpal testing

answered Jun 12 '13 at 01:36

simpletron

739
4
14

score 0 · Answer 4 · answered Jun 12 '13 at 01:25

0

Use a quantifier:

-\w+|\w+\s|(?:\w+\.+)+\w+\s
           ^^^      ^^

You can also simply your expression to:

-?\w+\s?|(?:\w+\.+)+\w+\s

answered Jun 12 '13 at 01:25

Andrew Cheong

29,362
15
90
145

score 0 · Answer 5 · answered Jun 12 '13 at 01:32

0

For doing this in java, all you need to do is split it along the spaces, no regex needed. The good ole String.split() should be able to handle it.

answered Jun 12 '13 at 01:32

AJMansfield

4,039
3
29
50

Regular expression - filename with multiple periods

5 Answers5