| regular expressions (parsing strings) [message #36958] |
Wed, 12 November 2003 12:04  |
setthivoine you
Messages: 1
Registered: November 2003
|
Junior Member |
|
|
Hi,
My IDL programs need to parse a text file which contains variables that
would need replacing at read-time. Specifically:
#os =unix
#windows_var1 =g:
#windows_var2 =/rootdir/
#unix_var1 =/mnt/groupserver/
#unix_var2 =/rootdir/
#var3 =< <os>_var1 >< <os>_var2 >data/
#var4 =<var3>counterfile.txt
So #var3 eventually becomes '/mnt/groupserver/rootdir/data/' and #var4
becomes 'mnt/groupserver/rootdir/data/counterfile.txt'. And if #os was
set to 'windows', #var4 becomes 'g:/rootdir/data/counterfile.txt'.
I am trying to use regular expressions to replace the text (specifically
using strepex.pro from
http://astro.uni-tuebingen.de/software/idl/aitlib/misc/strep ex.html )
but am having problems with nested tags.
Could anyone point me to somewhere that could help me out ?
Thanks
--Sett
|
|
|
|
| Re: Regular expression [message #53859 is a reply to message #36958] |
Fri, 04 May 2007 11:13  |
Foldy Lajos
Messages: 268
Registered: October 2001
|
Senior Member |
|
|
On Fri, 4 May 2007, kuyper@wizard.net wrote:
> FÖLDY Lajos wrote:
>> On Fri, 4 May 2007, Lasse Clausen wrote:
>>
>>> Hi there,
>>>
>>> why does
>>>
>>> print, stregex('[', '[\[]')
>>>
>>> work, i.e. produce 0, whereas
>>>
>>
>> You are searching for \ or [ ==> found.
> ...
>> \ loses its 'escape char' meaning in a bracket expression, and becomes an
>> ordinary character.
>
> In other cases, such as the unix vi command, the regular expression \
> [[^\]]*] matches any string that starts with '[', has an arbitrarily
> long string of characters that are not ']', followed by a ']'
> character. In IDL, however, stregex("ab[23]", "\[[^\]]*]") returns -1.
> Is there any simple way to perform a similar search using IDL regular
> expression?
>
It's easy, just omit the backslash: print, stregex("ab[23]", "\[[^]]*]")
If you want to put a ] in the non-matching list, put it right after the ^.
regards,
lajos
|
|
|
|
|
|
| Re: Regular expression [message #53863 is a reply to message #36958] |
Fri, 04 May 2007 09:36 |
lasse
Messages: 48
Registered: February 2007
|
Member |
|
|
On 4 May, 16:56, Allan Whiteford
<allan.rem...@phys.remove.strath.ac.remove.uk> wrote:
> Lasse,
>
> Either:
>
> regex='[a-zA-Z0-9]+\[[0-9]+\]'
>
> or:
>
> regex='[a-zA-Z0-9]{2}\[[0-9]{2}\]'
>
> depending on whether your 'bb' and '23' need to be exactly two
> characters long or not.
>
> Note also you may want to check whether you're matching a substring
> inside your search string or the complete string. I'm not sure what you
> want to do.
>
> Thanks,
>
> Allan
>
> Lasse Clausen wrote:
>> On 4 May, 16:21, FÖLDY Lajos <f...@rmki.kfki.hu> wrote:
>
>>> On Fri, 4 May 2007, Lasse Clausen wrote:
>
>>>> Hi there,
>
>>>> why does
>
>>>> print, stregex('[', '[\[]')
>
>>>> work, i.e. produce 0, whereas
>
>>> You are searching for \ or [ ==> found.
>
>>>> print, stregex(']', '[\]]')
>
>>>> prints -1?
>
>>> You are searching for \ followed by ] ==> not found.
>
>>>> print, stregex(']', '\]')
>
>>>> works (i.e. prints 0).
>
>>> You are searching for ] ==> found.
>
>>> \ loses its 'escape char' meaning in a bracket expression, and becomes an
>>> ordinary character.
>
>>> regards,
>>> lajos
>
>> mhmm, don't understand. Ok, here we go: I have a string like this
>
>> bb[23]
>
>> where bb can be any combination of alphanumerics and the number can be
>> anything. I am looking for the regular expression that will match the
>> whole thing. My first idea was (at the moment I am not bothered about
>> the order of the different parts):
>
>> regex = '[a-zA-Z0-9\[\]]+'
>
>> but alas!
>
>> print, stregex('bb[23]', regex)
>> 4
>
>> What?! And any combination of omitting or changing the \ character
>> will result in either IDL complainign about non-balanced brackets, a
>> match at position 4 or it won't match.
>
>> Help?
>
>> Cheers
>> Lasse
Thanks for the reply. I realized that I could do it the way you
(Allan) proposed, without including the brackets in the character
group, but I was being more academic and looking for a way to include
them in the character group. The following works
print, stregex('bb[23]', '[][0-9a-b]+', length=length) & print, length
0
6
however, order is, not surprisingly, essential:
print, stregex('bb[23]', '[[]0-9a-b]+', length=length) & print, length
-1
-1
Cheers
Lasse
|
|
|
|
| Re: Regular expression [message #53864 is a reply to message #36958] |
Fri, 04 May 2007 09:04 |
Foldy Lajos
Messages: 268
Registered: October 2001
|
Senior Member |
|
|
On Fri, 4 May 2007, Lasse Clausen wrote:
> mhmm, don't understand. Ok, here we go: I have a string like this
>
> bb[23]
>
> where bb can be any combination of alphanumerics and the number can be
> anything. I am looking for the regular expression that will match the
> whole thing. My first idea was (at the moment I am not bothered about
> the order of the different parts):
>
> regex = '[a-zA-Z0-9\[\]]+'
>
This regexp searches for a bracket expression (a-zA-Z0-9\[\) followed by
one or more ]'s. (\ behaves as an ordinary character after the opening
bracket [, so the first ] is the closing bracket.)
> but alas!
>
> print, stregex('bb[23]', regex)
> 4
>
3 matches the bracket expr. and ] matches itself. So the answer is 4.
> What?! And any combination of omitting or changing the \ character
> will result in either IDL complainign about non-balanced brackets, a
> match at position 4 or it won't match.
>
Try something like this:
[a-zA-Z0-9]+ one ore more alphanumeric char
\[ [
[0-9]+ one or more digits
] }
ie:
regex = '[a-zA-Z0-9]+\[[0-9]+]'
regards,
lajos
|
|
|
|
| Re: Regular expression [message #53867 is a reply to message #36958] |
Fri, 04 May 2007 08:56 |
Allan Whiteford
Messages: 117
Registered: June 2006
|
Senior Member |
|
|
Lasse,
Either:
regex='[a-zA-Z0-9]+\[[0-9]+\]'
or:
regex='[a-zA-Z0-9]{2}\[[0-9]{2}\]'
depending on whether your 'bb' and '23' need to be exactly two
characters long or not.
Note also you may want to check whether you're matching a substring
inside your search string or the complete string. I'm not sure what you
want to do.
Thanks,
Allan
Lasse Clausen wrote:
> On 4 May, 16:21, F�LDY Lajos <f...@rmki.kfki.hu> wrote:
>
>> On Fri, 4 May 2007, Lasse Clausen wrote:
>>
>>> Hi there,
>>
>>> why does
>>
>>> print, stregex('[', '[\[]')
>>
>>> work, i.e. produce 0, whereas
>>
>> You are searching for \ or [ ==> found.
>>
>>
>>> print, stregex(']', '[\]]')
>>
>>> prints -1?
>>
>> You are searching for \ followed by ] ==> not found.
>>
>>
>>> print, stregex(']', '\]')
>>
>>> works (i.e. prints 0).
>>
>> You are searching for ] ==> found.
>>
>> \ loses its 'escape char' meaning in a bracket expression, and becomes an
>> ordinary character.
>>
>> regards,
>> lajos
>
>
> mhmm, don't understand. Ok, here we go: I have a string like this
>
> bb[23]
>
> where bb can be any combination of alphanumerics and the number can be
> anything. I am looking for the regular expression that will match the
> whole thing. My first idea was (at the moment I am not bothered about
> the order of the different parts):
>
> regex = '[a-zA-Z0-9\[\]]+'
>
> but alas!
>
> print, stregex('bb[23]', regex)
> 4
>
> What?! And any combination of omitting or changing the \ character
> will result in either IDL complainign about non-balanced brackets, a
> match at position 4 or it won't match.
>
> Help?
>
> Cheers
> Lasse
>
|
|
|
|
| Re: Regular expression [message #53869 is a reply to message #36958] |
Fri, 04 May 2007 08:46 |
lasse
Messages: 48
Registered: February 2007
|
Member |
|
|
On 4 May, 16:21, FÖLDY Lajos <f...@rmki.kfki.hu> wrote:
> On Fri, 4 May 2007, Lasse Clausen wrote:
>> Hi there,
>
>> why does
>
>> print, stregex('[', '[\[]')
>
>> work, i.e. produce 0, whereas
>
> You are searching for \ or [ ==> found.
>
>> print, stregex(']', '[\]]')
>
>> prints -1?
>
> You are searching for \ followed by ] ==> not found.
>
>> print, stregex(']', '\]')
>
>> works (i.e. prints 0).
>
> You are searching for ] ==> found.
>
> \ loses its 'escape char' meaning in a bracket expression, and becomes an
> ordinary character.
>
> regards,
> lajos
mhmm, don't understand. Ok, here we go: I have a string like this
bb[23]
where bb can be any combination of alphanumerics and the number can be
anything. I am looking for the regular expression that will match the
whole thing. My first idea was (at the moment I am not bothered about
the order of the different parts):
regex = '[a-zA-Z0-9\[\]]+'
but alas!
print, stregex('bb[23]', regex)
4
What?! And any combination of omitting or changing the \ character
will result in either IDL complainign about non-balanced brackets, a
match at position 4 or it won't match.
Help?
Cheers
Lasse
|
|
|
|
|
|
| Re: Regular expression [message #53872 is a reply to message #36958] |
Fri, 04 May 2007 08:21 |
Foldy Lajos
Messages: 268
Registered: October 2001
|
Senior Member |
|
|
On Fri, 4 May 2007, Lasse Clausen wrote:
> Hi there,
>
> why does
>
> print, stregex('[', '[\[]')
>
> work, i.e. produce 0, whereas
>
You are searching for \ or [ ==> found.
> print, stregex(']', '[\]]')
>
> prints -1?
>
You are searching for \ followed by ] ==> not found.
> print, stregex(']', '\]')
>
> works (i.e. prints 0).
>
You are searching for ] ==> found.
\ loses its 'escape char' meaning in a bracket expression, and becomes an
ordinary character.
regards,
lajos
|
|
|
|
| Re: Regular expression [message #53927 is a reply to message #53863] |
Tue, 08 May 2007 05:10 |
Allan Whiteford
Messages: 117
Registered: June 2006
|
Senior Member |
|
|
Lasse Clausen wrote:
> On 4 May, 16:56, Allan Whiteford
> <allan.rem...@phys.remove.strath.ac.remove.uk> wrote:
>
>> Lasse,
>>
>> Either:
>>
>> regex='[a-zA-Z0-9]+\[[0-9]+\]'
>>
>> or:
>>
>> regex='[a-zA-Z0-9]{2}\[[0-9]{2}\]'
>>
>> depending on whether your 'bb' and '23' need to be exactly two
>> characters long or not.
>>
>> Note also you may want to check whether you're matching a substring
>> inside your search string or the complete string. I'm not sure what you
>> want to do.
>>
>> Thanks,
>>
>> Allan
>>
>> Lasse Clausen wrote:
>>
>>> On 4 May, 16:21, F�LDY Lajos <f...@rmki.kfki.hu> wrote:
>>
>>>> On Fri, 4 May 2007, Lasse Clausen wrote:
>>
>>>> >Hi there,
>>
>>>> >why does
>>
>>>> >print, stregex('[', '[\[]')
>>
>>>> >work, i.e. produce 0, whereas
>>
>>>> You are searching for \ or [ ==> found.
>>
>>>> >print, stregex(']', '[\]]')
>>
>>>> >prints -1?
>>
>>>> You are searching for \ followed by ] ==> not found.
>>
>>>> >print, stregex(']', '\]')
>>
>>>> >works (i.e. prints 0).
>>
>>>> You are searching for ] ==> found.
>>
>>>> \ loses its 'escape char' meaning in a bracket expression, and becomes an
>>>> ordinary character.
>>
>>>> regards,
>>>> lajos
>>
>>> mhmm, don't understand. Ok, here we go: I have a string like this
>>
>>> bb[23]
>>
>>> where bb can be any combination of alphanumerics and the number can be
>>> anything. I am looking for the regular expression that will match the
>>> whole thing. My first idea was (at the moment I am not bothered about
>>> the order of the different parts):
>>
>>> regex = '[a-zA-Z0-9\[\]]+'
>>
>>> but alas!
>>
>>> print, stregex('bb[23]', regex)
>>> 4
>>
>>> What?! And any combination of omitting or changing the \ character
>>> will result in either IDL complainign about non-balanced brackets, a
>>> match at position 4 or it won't match.
>>
>>> Help?
>>
>>> Cheers
>>> Lasse
>
>
> Thanks for the reply. I realized that I could do it the way you
> (Allan) proposed, without including the brackets in the character
> group, but I was being more academic and looking for a way to include
> them in the character group. The following works
>
> print, stregex('bb[23]', '[][0-9a-b]+', length=length) & print, length
> 0
> 6
>
> however, order is, not surprisingly, essential:
>
> print, stregex('bb[23]', '[[]0-9a-b]+', length=length) & print, length
> -1
> -1
>
> Cheers
> Lasse
>
Lasse,
That regular expression will pretty much match anything though:
IDL> print, stregex('bb[23]', '[][0-9a-b]+', length=length) & print, length
0
6
IDL> print, stregex('bba23a', '[][0-9a-b]+', length=length) & print, length
0
6
You can't put the square brackets in the range of characters to match
unless you're willing for them to be optional which I'd presume you
don't want. In the example above an 'a' is just as good as a '[' or a ']'.
Thanks,
Allan
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
