Steve,
The argument I wish to pass is either one string, or a list of strings, or a tuple of strings.
For instance, I have:
def abc(arg1, arg2, arg3)
Let say that I expect arg1 and arg3 to be a number, and arg2 can be either one string, or a bunch of strings and I need to do something on each of the strings passed down. So, using the isinstance function suggested, I can have:
def abc(arg1, arg2, arg3) if isinstance(arg2,(list,tuple)): for item in arg2: abc(arg1, item) else: # do whatever with arg2 and I know now arg2 is a single instance of a string ...
This way, when I invoke the function, I don't have to always remember the [] like:
abc(1,["String 1",],5)
I can simply do:
abc(1,"String 1",5)
and likewise, I can do:
abc(1,["String 1","String 2"],5)
Am I on the right track?
Yeah, given those constraints, you basically have two options -- check for list/tuple (as you have above) or check for str/unicode (as I do below). I personally prefer the latter because it seems much more likely that users will define alternate collection classes than that users will define alternate string classes. My code would look like:
def abc(arg1, arg2, arg3):
if not isinstance(arg2, basestring):
for item in arg2:
abc(arg1, item, arg3)
else:
# do whatever with arg2 (guaranteed to be a string)That way if you later call your code with, say:
abc(1, set(["String 1", "String 2"]), 5)
or
abc(1, collections.deque(["String 1", "String 2"]), 5)
it still works.
Steve
P.S. My real preference here would be to change the order of your arguments and write the function as:
def abc(arg1, arg3, *arg2s)
for arg2 in arg2s:
# do whatever with arg2 (should be a string)and then write your function calls as:
abc(1, 5, "String 1")
abc(1, 5, "String 1", "String 2")Then you don't ever have to remember []s. ;-) But if you really are stuck with arg2 as the second argument, I'd probably go for testing isinstance(x, basestring).
Steve -- http://mail.python.org/mailman/listinfo/python-list
