Using .values.max() for arrays

Question

I use the following code to get a row from a dataframe and then find the max value.

def find_max(a):
    return a.values.max()

row = df.iloc[0].astype(int)
max_value = find_max(a)

That works fine. However, if I pass an array like

ar = [1,2,3]
max_value = find_max(ar)

it doesn't work and I receive AttributeError: 'list' object has no attribute 'values'. How can I use that function for both types?

something along the lines of isinstance dataframe else max()? — Bharel
– Bharel, Commented Jan 26, 2022 at 13:24
Note: a Python list is not generally referred to as an array. Pandas builds on top of numpy, which does have array type. You could use numpy.max(), and test if .values exists. — Martijn Pieters
– Martijn Pieters, Commented Jan 26, 2022 at 13:24

Martijn Pieters · Accepted Answer · 2022-01-26 13:30:12Z

6

You'll have to test for the type, or use exception handling, to switch approaches:

def find_max(a):
    if isinstance(a, list):
        return max(list)
    return a.values.max()

or

def find_max(a):
    try:
        return a.values.max()
    except AttributeError:
        # assume a Python sequence
        return max(a)

Which one you pick depends on a few factors:

Do you need to support other iterables, like tuples or sets or generators?
What is the more common case; the try...except case is faster if the majority of calls pass in a Pandas series.

Another option is to use np.max(), which works on either type:

import numpy as np

def find_max(a):
    return np.max(a)

Note that there is no need to use the .values attribute here!

edited Jan 26, 2022 at 13:30

answered Jan 26, 2022 at 13:26

Martijn Pieters

1.1m326 gold badges4.2k silver badges3.4k bronze badges

Sign up to request clarification or add additional context in comments.

4 Comments

Bharel Over a year ago

I would have actually checked for a dataframe type instead of a list, as max "covers" everything but that, but I like the EAFP approach better anyway

Martijn Pieters Over a year ago

@Bharel: it depends. What other types are you passing in?

Bharel Over a year ago

I'm not the OP, but considering .values.max() will work only for specific types while max is for general iterables, I'd special case the first.

Martijn Pieters Over a year ago

@Bharel: I meant the general 'you' more than the specific. :-)

Jafar Isbarov · Accepted Answer · 2022-01-26 13:33:21Z

5

def find_max(a):
    if isinstance(a, list):
        return max(a)
    elif isinstance(a, pd.DataFrame):
        return a.values.max()
    else:
        print("No max method found for the input type")

edited Jan 26, 2022 at 13:33

answered Jan 26, 2022 at 13:24

Jafar Isbarov

1,5822 gold badges12 silver badges28 bronze badges

3 Comments

Jafar Isbarov Over a year ago

I assumed that DataFrame is the default input type. How would you improve it?

Patrick Artner Over a year ago

Its fine hence the edit and upvote - try: except: would work for any sequence but it is not needed

Jafar Isbarov Over a year ago

@PatrickArtner Thanks. How does a try-except block compare to my current (edited) solution? Would that be better?

Robin · Accepted Answer · 2022-01-26 13:33:07Z

4

Why don't try numpy's integration for that?

import numpy as np

row = df.iloc[0].astype(int)
max_value = np.max(a.values)

ar = [1,2,3]
max_value = np.max(ar)

This way, you don't have to define a function. In general it`s more pythonic to use already existing and well tested functionality from python packages.

answered Jan 26, 2022 at 13:33

Robin

1456 bronze badges

Collectives™ on Stack Overflow

Using .values.max() for arrays

3 Answers 3

4 Comments

3 Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

4 Comments

3 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related