Le Tue, 3 Feb 2009 17:36:41 -0500, Kent Johnson <ken...@tds.net> a écrit :
> I think your use of the csv module is fine. What I really meant to say > was more like, I would have looked to pyparsing to solve the same > problem, and if you want a parser that parses the file into meaningful > records, then it might still be worth a look. This would require heavy use of the setResultsName() method to attach "semantic tags" to the parse results; so that when you dig into them you know what each snippet is -- without the need of partial reparsing ;-) The point can be illustrated with the simple case of parsing arithmetic operations. Imagine each operand of a '+' can be a (litteral) number, a symbol (name), a grouped (parenthesized) sub-operation, an operation of higher priority. The result beeing of the form [op1, '+', op2] There is no other way to know what kind of thing op1 and op2 are, if they don't themselves carry the information, than re-examining them. Which has already beeing done while parsing. (The information needed is the name of the parsing rule used to yield the result.) Denis > > Kent ------ la vida e estranya _______________________________________________ Tutor maillist - Tutor@python.org http://mail.python.org/mailman/listinfo/tutor