[R] unexpected behaviour of sub() / usage of regexp

[Prof Brian Ripley]

On 09/12/2011 14:49, Jannis wrote:
> Thanks to all who replied. perl = TRUE indeed seems to fix the problem. It would be great, however, to prevent others from stumbling in this pitfall by fixing the issue if this is possible. But as Prof. Ripley mentioned fixing this might be difficult/impossible so we might have to live with it.
>
>
> By the way, is there an easily accessible and search able list of such bugs for R (just for the future)?

http://www.bugs.r-project.org

I'm not sure how obvious it would be that it is the same problem.  I 
happened to have worked on trying to solve it.

>
> Thanks a lot
> Jannis
>
>
>
> ----- Ursprüngliche Message -----
> Von: Sarah Goslee<sarah.goslee at gmail.com>
> An: Duncan Murdoch<murdoch.duncan at gmail.com>
> Cc: Jannis<bt_jannis at yahoo.de>; "r-help at r-project.org"<r-help at r-project.org>
> Gesendet: 15:37 Freitag, 9.Dezember 2011
> Betreff: Re: [R] unexpected behaviour of sub() / usage of regexp
>
> But I do get the incorrect result on R 2.14.0 on linux:
>> sub('[[:digit:]]{1,2}', '', '9ewww')
> [1] "www"
>
> And also:
>
>> sub('[[:digit:]]{1,2}', '', '9ewww')
> [1] "www"
>> sub('[[:digit:]]{1,2}', '', 'ewww9')
> [1] "ww9"
>> sub('\\d{1,2}', '', 'ewww9')
> [1] "ww9"
>
> But:
>> sub('\\d', '', 'ewww9')
> [1] "ewww"
>> sub('\\d*', '', '9ewww')
> [1] "ewww"
>
> So it seems to be something about the way the curly braces are
> handled, but only with certain groups:
>
>> sub('e{1,2}', '', '9ewww')
> [1] "9www"
>> sub('9{1,2}', '', '9ewww')
> [1] "ewww"
>
>
> But, as Prof. Ripley's email suggests, perl=TRUE solves the problem.
> (I was trying out various combinations when it appeared in my inbox.)
>
>> sessionInfo()
> R version 2.14.0 (2011-10-31)
> Platform: x86_64-redhat-linux-gnu (64-bit)
>
> locale:
> [1] LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C
> [3] LC_TIME=en_US.UTF-8        LC_COLLATE=en_US.UTF-8
> [5] LC_MONETARY=en_US.UTF-8    LC_MESSAGES=en_US.UTF-8
> [7] LC_PAPER=C                 LC_NAME=C
> [9] LC_ADDRESS=C               LC_TELEPHONE=C
> [11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C
>
> attached base packages:
> [1] stats     graphics  grDevices utils     datasets  methods   base
>
>
>
> On Fri, Dec 9, 2011 at 9:25 AM, Duncan Murdoch<murdoch.duncan at gmail.com>  wrote:
>> On 09/12/2011 9:20 AM, Jannis wrote:
>>>
>>> Dear R users,
>>>
>>>
>>> the way I understand the documentation of sub() and regexp the following
>>> code:
>>>
>>>
>>>
>>> sub('[[:digit:]]{1,2}', '', '9ewww')
>>>
>>>
>>>
>>> ... should yield:
>>>
>>> 'ewww'
>>>
>>>
>>> It returns, however:
>>>
>>> 'www'
>>>
>>>
>>> Why is this the case? My code should just substitute 1 (minimum) or up to
>>> 2 (maximum) digits, i.e. numbers and not the 'e' in the string. Do I
>>> misinterpret something here?
>>
>>
>> I get your expected output of "ewww" running 2.14.0 or 2.14.0-patched on
>> Windows.   So it's not a universal problem...
>>
>> Duncan Murdoch
>>
>>>
>>> Thanks for any ideas
>>> Jannis
>>>
>>>
>>>>   sessionInfo()
>>> R version 2.14.0 (2011-10-31)
>>> Platform: i686-pc-linux-gnu (32-bit)
>>>
>>> locale:
>>>   [1] LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C                [3]
>>> LC_TIME=en_US.UTF-8        LC_COLLATE=en_US.UTF-8      [5]
>>> LC_MONETARY=en_US.UTF-8    LC_MESSAGES=en_US.UTF-8     [7] LC_PAPER=C
>>>            LC_NAME=C                   [9] LC_ADDRESS=C
>>> LC_TELEPHONE=C            [11] LC_MEASUREMENT=en_US.UTF-8
>>> LC_IDENTIFICATION=C
>>> attached base packages:
>>> [1] stats     graphics  grDevices utils     datasets  methods   base
>>>
>


-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595