2Q's: How to autocreate instance of class;How to check formembership in a class

[asdf]

So I'm writing a script which will create several instances of User()
class. I want each instance to be named after the login name
of a user. I don't know beforehand how many users the script will
have to create or how they are named. Right now I've created a dictionary
of usernames as keys and objects themselves as values. It works,
but is very unwieldy, because every time I need to access a value from
within an object I have to do something like dict-user[x].value.
Is there a way of automatically naming objects from variable names.
So for example if I know that var1=jsmith. Can I somehow do
var1=User(). This obviously won't work because I tried this. It'll just
create var1 of type User.

My second question is how can I check if object is a member of a class.
so let's say I create a=User(), b=User()...
Can I do something similar to
if x.Users()==TRUE:
print "user already created"

Right now I'm doing this using try-except which works
but I'm looking for something cleaner.

thanks for all the replies.

[George Sakkis]

On Jun 16, 9:16*pm, asdf <[e-mail]:

So I'm writing a script which will create several instances of User()
class. I want each instance to be named after the login name
of a user. I don't know beforehand how many users the script will
have to create or how they are named. Right now I've created a dictionary
of usernames as keys and objects themselves as values.

That's probably the most reasonable representation.

It works,

but is very unwieldy, because every time I need to access a value from
within an object I have to do something like dict-user[x].value.

It seems you are not aware of being able to name intermediate objects:

# name2user: mapping of names to User instances
for name in 'Bob', 'John', 'Sue':
u = name2user[name]
print u.value
print u.password
...

In the sample above, u refers to whatever name2user[name] refers
(hopefully a User instance here), it does not create a copy. Therefore
after the assignment you can refer to the user as u instead of
name2user[name]. Not only this is easier to read but, unlike
statically typed languages, it is faster too. An expression such as
"a[b].c[d]" involves (at least) three method calls, so as long as it
doesn't change it is faster to compute it once, give it a name and
refer to the name afterwards.

My second question is how can I check if object is a member of a class.
so let's say I create a=User(), b=User()...
Can I do something similar to
if x.Users()==TRUE:
* * * * print "user already created"

I am not sure what you mean here. If you want to check whether a user
with a given name exists, look it up in the dictionary:

if 'Bob' in name2user:
...

If you ask how to check if some name x is a User instance, use the
isinstance() function:

if isinstance(x, User):
print x.password

Checking for class membership though is usually considered unnecessary
or even harmful in Python. Just assume 'x' is a User and use it
anyway.

George

[Mark Wooding]

asdf <as**@asdf.comwrote:

(Presumably nothing to do with the Common Lisp system-definition utility.)

So for example if I know that var1=jsmith. Can I somehow do
var1=User().

Something like this might work.

class User (object):
def __init__(me, name):
me.name = name

class Users (object):
def __getattr__(me, name):
try:
return object.__getattr__(me, name)
except AttributeError:
u = me.__dict__[name] = User(name)
return u

>>users = Users()
alice = users.alice
alice.name

'alice'

>>alice is users.alice

True

Not very nice, though, and not particularly good at defending against
typos.

My second question is how can I check if object is a member of a class.
so let's say I create a=User(), b=User()...

The built-in isinstance function seems an obvious choice.

-- [mdw]