Re: [Pyparsing] Grouping when using asList()

[Paul M. <pt...@au...>]

Joshua -

<lightbulb>Ah, now I see why you are confused!</lightbulb>

The Group class does *not* do grouping like you see in SQL GROUP BY or the
itertools groupby.  Group is a way for a grammar developer to impart
structure to the parsed tokens.  The default behavior is for all the tokens
to be returned in one flat list.  Here is an example:
    
testData = "123 abc 234 def 789 456 xyz"
entry = Word(nums) + Optional(Word(alphas))
grammar = OneOrMore(entry)
print grammar.parseString(testData).dump()

Prints:
['123', 'abc', '234', 'def', '789', '456', 'xyz']

Now we will use Group to "group" the tokens for each entry:
    
entry = Group( Word(nums) + Optional(Word(alphas)) )

Prints:
[['123', 'abc'], ['234', 'def'], ['789'], ['456', 'xyz']]


You are looking for something a little different.  Pyparsing has a feature
to implement it, but I've not seen it get much use.  It is a qualifier on
setResultsName.  Before I describe the qualifier itself, let me again show
you the default behavior.

entry = Word(nums).setResultsName("int") +
Optional(Word(alphas).setResultsName("word"))

prints:
['123', 'abc', '234', 'def', '789', '456', 'xyz']
- int: 456
- word: xyz

The last matching string is the one that is saved for the given results
name.  But sometimes, you want to keep *all* the matching strings.  This is
done using the listAllMatches qualifier:
    
entry = Word(nums).setResultsName("int",listAllMatches=True) +
Optional(Word(alphas).setResultsName("word",listAllMatches=True))

prints:
['123', 'abc', '234', 'def', '789', '456', 'xyz']
- int: ['123', '234', '789', '456']
- word: ['abc', 'def', 'xyz']

That's it for now, *this* is a long e-mail! :)

I'll post the mods to your code in the next e-mail.

-- Paul