Re: Backslashes in regular expression replacement patterns

[Juan Pablo Carbajal]

On Wed, Oct 17, 2012 at 6:31 PM, Rik <address@hidden> wrote:
> On 10/16/2012 02:37 PM, Ben Abbott wrote:
>> On Oct 16, 2012, at 5:00 PM, Rik wrote:
>>
>>> On 10/16/2012 01:28 PM, Michael D. Godfrey wrote:
>>>> On 10/16/2012 12:29 PM, Rik wrote:
>>>>> regexprep ('a', '(\w)', "\\$1")
>>>>>
>>>>> Should it be '\a' or '$1'?
>>>>>
>>>>> In the first case, one would get the same result from using a replacement
>>>>> pattern of '\\$1' or "\\$1" which is odd because usually single quotes and
>>>>> double quotes are not interchangeable.  On the other hand, Perl would
>>>>> return '\a' for the regexprep above.
>>>>>
>>>> Matlab R2009b returns '\a'
>>>>
>>> Matlab has support for double-quoted strings now?  I thought they didn't 
>>> support them, and this was only an Octave behavior question.
>>>
>>> --Rik
>> No.  Only single-quoted strings
>>
>> regexprep ('a', '(\w)', '\\$1')
>>
>> ans =
>>
>> \a
>
> 10/17/12
>
> All,
>
> I have single-quoted strings working.  My real issue is double-quoted
> strings which are specific to Octave.  With an expression like "\$1" the
> interpreter itself does one round of escape sequence processing and changes
> this to "$1".  So when I get the replacement string in regexprep I already
> can't tell whether the user typed "\$1" or "$1".  I imagine making a
> special case in the parser to not do double quote escape processing for a
> single function might be difficult (any ideas jwe?).  If that is the case
> then I think I will keep the current behavior of two rounds of expansion
> and just document it.
>
> --Rik

Hi Rik,

"\$1" is an unrecognized escape sequence. You get a warning.
"\\$1" produces the same as '\$1' which looks correct to me.