[Date Prev][Date Next][Thread Prev][Thread Next]
[Date Index]
[Thread Index]
- Subject: Re: Lost (again) with strings patterns
- From: Ico Doornekamp <lua@...>
- Date: Sat, 16 Nov 2013 20:02:34 +0100
* On 2013-11-16 19:42:36 +0100, Laurent FAILLIE wrote:
> 2 questions related to the same kind of strings :
>
> 1/ I have a string like this one :
> l="Débit ATM 3939 kb/s 945 kb/s"
>
> How can I extract only numbers ?
Make sure the pattern expects the characters between the numbers:
l:match("(%d+).-(%d+)")
> 2/ If I want to generalize, I got a set of strings link
>
> Débit ATM 3939 kb/s 945 kb/s
> Marge de bruit 8.10 dB 8.20 dB
> Atténuation 56.00 dB 33.40 dB
> FEC 21976 196
> CRC 600 0
> HEC 38 184
>
> I would like to retrieve :
> lab : the line label ('Débit ATM' or 'Marge de bruit', ...)
> val1 : first value
> val2 : second value.
Assuming there is always two or more spaces or tabs between the labels
and the rest of the line:
label, val1, val2 = l:match("(.-)%s%s+([%d%.]+).-([%d%.]+)")
(.-) Match as few characters as possible ...
%s%s+ ... up to two or more spaces
([%d%.]+) Match a sequence of digits and dots ...
.- ignore stuff in between ...
([%d%.]+) ... and again match a sequence of digits and dots
Beware: in your original mail, as received by my MUA, the white space
between the labels and the numbers are not encoded by spaces or tabs,
but by ascii 0xa0 (non breaking space), which is not matched by the %s
in the patterns.
This might be caused by the transit by mail, but if your original data
contains nbsp's as well, the above pattern will fail.
--
:wq
^X^Cy^K^X^C^C^C^C