pyparsing seems to treat numbers differently than chars?
Brought to you by:
ptmcg
Menu
▾
▴
(using: pyparsing 2.1.0 (also tried 2.0.7) and Python 3.5.0)
In playing around with pyparsing in preparation trying to use it, I started running into some weird cases where it seemed to work incorrectly, after trying some different approaches, I ended up finding that, at least for me, using numbers in Word() seems to operate differently than passing letters… for example:
Example that started it, I was trying to parse numbers from 10 to 59
So, then I tried it with text from another example, which worked
>>> a1 = Word('abcde', max=1) >>> a2 = Word('fghij', max=1) >>> b = a1 + a2 >>> b.parseString('af') (['a', 'f'], {})And tried combining text and numbers, which worked:
>>> b1 = Word('abcde12345') >>> b2 = Word('jklmn67890') >>> c = b1 + b2 >>> c.parseString('18') (['1', '8'], {})So, then I tried the variable "nums" (maybe I was structuring my strings wrong)?
(OK, I know that the below would work better like: d = Word(nums, max=2), but I was trying to duplicate my problem<grin>)
>>> d = Word(nums, max=1) + Word(nums, max=1) >>> d.parseString('12') Traceback (most recent call last): File "<input>", line 1, in <module> File "C:\Users\dstrohl\Documents\PyCharm Projects\VEnvs\is_email\lib\site-packages\pyparsing.py", line 1129, in parseString raise exc File "C:\Users\dstrohl\Documents\PyCharm Projects\VEnvs\is_email\lib\site-packages\pyparsing.py", line 1119, in parseString loc, tokens = self._parse( instring, 0 ) File "C:\Users\dstrohl\Documents\PyCharm Projects\VEnvs\is_email\lib\site-packages\pyparsing.py", line 993, in _parseNoCache loc,tokens = self.parseImpl( instring, preloc, doActions ) File "C:\Users\dstrohl\Documents\PyCharm Projects\VEnvs\is_email\lib\site-packages\pyparsing.py", line 2390, in parseImpl loc, resultlist = self.exprs[0]._parse( instring, loc, doActions, callPreParse=False ) File "C:\Users\dstrohl\Documents\PyCharm Projects\VEnvs\is_email\lib\site-packages\pyparsing.py", line 997, in _parseNoCache loc,tokens = self.parseImpl( instring, preloc, doActions ) File "C:\Users\dstrohl\Documents\PyCharm Projects\VEnvs\is_email\lib\site-packages\pyparsing.py", line 1807, in parseImpl raise ParseException(instring, loc, self.errmsg, self) pyparsing.ParseException: Expected W:(0123...) (at char 1), (line:1, col:2)So... any thoughts? am I doing something wrong here?
Last edit: Dan Strohl 2016-02-12
Ok, I dont think it is a bug, I am doing something wrong... I just tried some more things:
w = Word('0123456789', max=1) x = Word('12345', max=1) y = x+w z = y.parseString('16') print(z) d = Word('0123456789', max=1) v = Word('12345', max=1) print(d.parseString('1')) print(v.parseString('2')) j = d+v t = x+w print(j.parseString('12')) print(t.parseString('12'))it worked for everything up to print(j.parseString('12')).
in playing around some more, I think that "y" worked becaues in "16", the "1" is in x, but not the "6", so it moved on... for "12", the "1" and "2" are both in x, but the max=1 is there, so it errored out....
my expectation was that it would try x, error out, then try y, not try x and error out without trying y... so... I am sure I am missing something fundemantal here, can you help?
Last edit: Dan Strohl 2016-02-12
aaannnddddd.... when I replace the "max=1" with "exact=1", it seems to work... so... bug or id(10).t error?
This seems the logical behavior to me - "max" meaning "no more than", and in parsing "12", you had "more than", so "fail". Whereas "exact" implies that "more than this is not necessarily bad, just parse this exact amount." (Sorry for the delay in responding, I don't visit this Discussion board very often.) Also, each expression in the grammar has to pass or fail on its own, it doesn't lookahead to the next expression to see if there should be some partial match in order to leave a valid bit for the next expression in the grammar. You can implement this kind of lookahead, using NotAny and FollowedBy expressions, but it has to be explicit.