Parsing configuration file using pyparsing

Question

I am trying to use pyparsing to parse a configuration file of following form

x = "/user/test"
y = 3

Here is my code snippet

ParserElement.defaultWhitespaceChars = (" \t")
END = StringEnd()
NL = LineEnd().suppress()
assignment = Literal('=')

key_str = CharsNotIn("=")
value_str = Group(~assignment + restOfLine)

line = Group(key_str + assignment + value_str)
lines = ZeroOrMore(line)
lines.ignore(NL)

text = """
y = 3
x = 2
"""

The output that I get from parseFile tells me it is parsing the first line only. Can anybody please help me find out what I am doing wrong?

Would you be averse to using a library that can already parse files that look sort of like this? For example pypi.python.org/pypi/configobj — SingleNegationElimination
– SingleNegationElimination, Commented May 9, 2013 at 17:43
Thank you for your input.There are additional parsing work I am doing (e.g. parsing of a scheduler log, some script files etc). In that context I am trying to get a better understanding of pyparsing for last few weeks. I consider myself a novice in pyparsing area so wanted to get some pointers about what I did wrong here. — Abhijit Bhattacharya
– Abhijit Bhattacharya, Commented May 9, 2013 at 17:51
@AbhijitBhattacharya - search SO for other questions/answers with the pyparsing tag. You can also visit the wiki at pyparsing.wikispace.com, and read through the Discussion tab on the wiki Home page. — PaulMcG
– PaulMcG, Commented May 10, 2013 at 13:41

Hooked · Accepted Answer · 2013-05-09 18:22:45Z

2

It looks like you are on the right track. Maybe you are doing something wrong when you acutally pass the text to the grammar. Adding the following line in your code

print lines.parseString(text)

Gives the expected output

[['y ', '=', [' 3']], ['x ', '=', [' 2']]]

As an aside, normally you don't want to keep the whitespace when parsing. The tokens are usually the only thing that matters. This is how I would parse your example:

EOL = LineEnd().suppress()
EQ  = Literal("=").suppress()
val = Word(alphanums)
line = Group(val('key') + EQ + val('val') + EOL)
grammar = OneOrMore(line)

for x in grammar.parseString(text):
    print x.dump()

The output in this case is much nicer

['y', '3']
- key: y
- val: 3
['x', '2']
- key: x
- val: 2

edited May 9, 2013 at 18:22

answered May 9, 2013 at 18:13

Hooked

88.9k46 gold badges197 silver badges271 bronze badges

Sign up to request clarification or add additional context in comments.

7 Comments

Hooked Over a year ago

No problem, and welcome to Stack Overflow! Typically the protocol is to upvote useful answers and "accept" an answer that solves your problem.

Hooked Over a year ago

I've updated the answer to show a different approach at the grammar.

Abhijit Bhattacharya Over a year ago

I accepted your answer. As for the upvote, I guess there is something called "reputation" for newcomers. Without that it doen't allow upvoting.

Hooked Over a year ago

@AbhijitBhattacharya I didn't know there was a rep limit for upvoting your own question's answers. According to the FAQ, stackoverflow.com/faq#reputation, you'll need at least 15 rep to do so... looks like you're halfway there!

PaulMcG Over a year ago

Try naming it "line*" - the trailing asterisk is a flag to pyparsing to keep all values, not just the last one; it is a shorthand for calling setResultsName with listAllMatches=True. Also, I overlooked one other point - the original question also had a RHS value of a quoted string as one of the examples. You can accommodate this by adding | quotedString to the definition of val.

|

Collectives™ on Stack Overflow

Parsing configuration file using pyparsing

1 Answer 1

7 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

7 Comments

Your Answer

Sign up or log in

Post as a guest

Related