Prev: Array error
Next: c++
From: consultmac2 on 17 Jan 2008 10:55
I am using C++ in VS 8 with the lovely boost libraries.

I am processing a log file where the useful information follows a
predictably formatted date/time stamp surrounded by [ ] characters.
A literal example of the text that I want to strip off is:
[ Tue Jan 15 14:30:59 2008 ]

a full example of a typical log line is:
[ Tue Jan 15 14:30:59 2008 ] Fast Food in T:\restaurants\london

As the result of my call to the boost function "regex_replace",
indicating that I want it to replace the match with an empty string, I
expect the result to be (in the above example's case):
Fast Food in T:\restaurants\london

The [ and ] characters have special meaning in regular expressions
and would seem to need to be 'escaped' somehow. In my program, I am
using the boost function: regex_replace, and tried the following in
the argument for the regular expression:
...., regex("\[ .* \] "), ...
The compiler gives me warnings that both '[' and ']' are "unrecognized
character escape sequence"s, but compiles successfully. However, the
result of running the regex_replace with that expression on the string
in the above example is:

[TueJan1514:30:592008]FastFoodinT:\restaurants\london

.... which is NOT what I expected. It is essentially just removing all
spaces from the text.

Thinking that perhaps the use of the '[' and ']' characters is not
functional in this case, I changed the argument for the regular
expression to:
...., regex("\x5B .\ \x5D "), ... which provides the hex
versions of the ASCII equivalents for the two symbols instead. With
that, the compilation warnings go away, but the results are exactly
the same as before (strips all blanks spaces from the text).

I take another stab and change the regular expression argument to:
...., regex("^ .* \] "), ... and get the one compiler warning
about the ']' character, but the results indicate that NOTHING gets
matched and removed from the text.

I MUST be missing something. Suggestions appreciated.

Thanks.
From: Mara Guida on 17 Jan 2008 14:06
consultmac2 wrote:
> ..., regex("\[ .* \] "), ...

regex("\\[ .* \\]")
From: Anand Hariharan on 17 Jan 2008 23:24
On Jan 17, 1:06 pm, Mara Guida <maragu...(a)gmail.com> wrote:
> consultmac2 wrote:
> > ..., regex("\[ .* \] "), ...
>
> regex("\\[ .* \\]")

Sadly enough, we'll have to wait until the next version of the
standard makes way into popular implementations, to get language
support for raw string literals (akin to Python).

http://groups.google.com/group/comp.lang.c++/msg/32a7144d59aba13b

- Anand
From: Jerry Coffin on 19 Jan 2008 01:34
In article <0a03199b-0eee-4635-913e-
5d3908cf201f(a)c4g2000hsg.googlegroups.com>, barryc.ctr(a)gmail.com says...
> I am using C++ in VS 8 with the lovely boost libraries.
>
> I am processing a log file where the useful information follows a
> predictably formatted date/time stamp surrounded by [ ] characters.
> A literal example of the text that I want to strip off is:
> [ Tue Jan 15 14:30:59 2008 ]
>
> a full example of a typical log line is:
> [ Tue Jan 15 14:30:59 2008 ] Fast Food in T:\restaurants\london
>
> As the result of my call to the boost function "regex_replace",
> indicating that I want it to replace the match with an empty string, I
> expect the result to be (in the above example's case):
> Fast Food in T:\restaurants\london

Unless your data has more variation than you've said, you don't need a
regex at all. To create a string without the timestamp, you could do
something like:

std::string stripped(input, input.find(']')+2, -1);

The "+2" part is to skip across the ']' and the space that follows.

--
Later,
Jerry.

The universe is a figment of its own imagination.
 | 
Pages: 1
Prev: Array error
Next: c++