Keyword argument parsing in error in 2.0?
thinkit
thinkit8 at lycos.com
Tue Jun 12 20:29:25 EDT 2001
i'd consider looking at this, but it contains a decimal number. use hexadecimal
only if you want to be taken seriously.
In article <mObU6.18156$i4.700435 at e420r-atl1.usenetserver.com>, "Steve says...
>
>I'm working on a SQL abstraction layer, using methods to generate SQL
>appropriate to a particular database back-end. This has involved using
>keyword arguments to many of the methods. As a catchall (to avoid having to
>handle too much SQL syntax) many methods have a REST keyword argument, which
>is appended to the generated statement. The 2.0 compiler is complaining
>about the following statement:
>
>c.UNION(c.SELECT(["l.beer", "l.drinker", ["count(distinct s.bar)",
>"nbars"]],
> FROM=[["likes", "l"], ["serves", "s"]],
> WHERE="l.beer=s.beer",
> REST="group by l.beer, l.drinker"),
> c.SELECT("distinct beer, drinker, 0 as nbars",
> FROM="likes",
> WHERE="beer not in (%s)" %
> c.SELECT("beer",
> FROM="serves")),
> REST="order by 3 desc")
>
>This validates fine in PythonWin, but when the module containing it is
>executed I get:
>
>Traceback (most recent call last):
> File "alltest.py", line 396, in ?
> REST="order by 3 desc"),
>TypeError: keyword parameter redefined: REST
>
>The implication is clearly that the system is seeing two definitions of the
>same keyword argument, but I have tried to analyze the statement carefully
>and it seems to me that the different REST arguments are being provided to
>different calls.
>
>Just to check, I tried running the following program:
>
>def a(x, y, z, REST="NothingA"):
> return "%s %s %s %s" % (x, y, z, REST)
>
>def b(x, y, z, REST="NothingB"):
> return "%s %s %s %s" % (x, y, z, REST)
>
>print "STARTING"
>
>print a("a",
> b("m", "n", "o", REST="Ignore1"),
> a("s", "t", "u", REST="Ignore2"),
> REST="Ignore3")
>
>This runs fine, printing
>
>STARTING
>a m n o Ignore1 s t u Ignore2 Ignore3
>
>Is this a bug, or can I simply not see my error for looking too hard?
>
>regards
> Steve
>
>
>
>
More information about the Python-list
mailing list