Python: return function variables and assign to new variables

Question

I have a function that returns a dict of pandas dataframes. The dict has a key and the key for each dataframe is its corresponding name.

def fun(inputs): x = calculations y = calculations z = calculations, etc. return: {'name1':x,'name2':y,'name3':z}

I would like to use some of these dataframes as variables elsewhere in my code. To do this, I currently do the following:

x = fun() var1 = x['name1'] var2 = x['name2'] var3 = x['name3']

and so on. I would like to do this more efficiently (fewer lines of code). In my mind, I should be able to do something like this:

[var1,var2,var3] = fun()

But this just assigns each name to each variable as a string. Can this be done? or should my function output be setup differently?

@juanpa.arrivillaga: Reaching for the [''] keys all the time is awkward, it makes your statements bulkier, and your IDE won't be as helpful with dict entries as with variables. — user2357112
– user2357112, Commented Apr 25, 2017 at 20:22
Then don't use a dictionary, and return a tuple and use tuple unpacking assignment. — juanpa.arrivillaga
– juanpa.arrivillaga, Commented Apr 25, 2017 at 20:24
@juanpa.arrivillaga: But then you have to remember which slots you put everything in, and you're stuck with whatever order you pick. namedtuples can help, but they still have issues. — user2357112
– user2357112, Commented Apr 25, 2017 at 20:29
... OK, I've never found it too difficult to keep it all organized, especially if you use namedtuples. But regardless, look at the functionality the OP is imagining utilizing this: [var1,var2,var3] = fun() how does this in any way take advantage of the nice properties of dictionaries we all know and love? — juanpa.arrivillaga
– juanpa.arrivillaga, Commented Apr 25, 2017 at 20:33

user2357112 · Accepted Answer · 2017-04-25 20:44:05Z

namedtuple

import the namedtuple function from collections
Use it to create a new class that you can then use to instantiate new class members
Accessing the named elements now utilizes the dot notation.

from collections import namedtuple mydfs = namedtuple('MyDFs', ['name1', 'name2', 'name3']) def fun(inputs): x = calculations y = calculations z = calculations, etc. return mydfs(x, y, z) x = fun(inputs) x.name1

If you want, you can also unpack one of these:

var1, var2, var3 = x

Old Answer

Easiest way is to pass that dictionary to the pd.Panel constructor

pn = pd.Panel({'name1':x,'name2':y,'name3':z})

This is can now be thought of as a series of dataframes. Each dataframe can be accessed via pn['name1'] etc.

So... we access the dataframes exactly the same way we would have with the dict? It's not clear how this gets us anywhere.
@user2357112 True. It's more convenient cohesive structure. But access is identical. I'll have to rethink it.
pd.Panel is also facing an oncoming depreciation, I recently learned.

Collectives™ on Stack Overflow

Python: return function variables and assign to new variables

1 Answer 1

5 Comments

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

5 Comments

Related