Error when looking up max value from a python dictionary : max() arg is an empty sequence

Question

I am trying to filter a dictionary and find the max from a list of values. Here a sample dictionary:

import numpy as np

def clean_currency(x):
""" If the value is a string, then remove currency symbol and delimiters
otherwise, the value is numeric and can be converted
"""
    if isinstance(x, str):
       return(x.replace('$', '').replace(',', ''))
    return(x)



d = {
     'shores': ['$0.00'],
     'Broderick': ['$0.00', '$0.00', '$0.00'],
     '100 Broderick': ['$0.00', '$1,142,070.00', '$0.00', '$0.00', '$0.00'],
     '1001 Orange Grove': ['$0.00'],
     '1008 Hyde': [np.nan]
    }


# Keep the highest value from the list. 

revd = {
        k: max((e for e in v if pd.notna(e)), key=lambda x: float(clean_currency(x)) if isinstance(x, str) else x)
        for k,v in d.items()
       }

revd

I am getting ValueError: max() arg is an empty sequence error.

Full traceback

ValueError                                Traceback (most recent call last)
/var/folders/d0/gnksqzwn2fn46fjgrkp6045c0000gn/T/ipykernel_49841/938463883.py in <module>
      1 # Keep the highest value from the list.
      2 
----> 3 revd = {
      4         k: max((e for e in v if pd.notna(e)), key=lambda x: float(clean_currency(x)) if isinstance(x, str) else x)
      5         for k,v in d.items()

/var/folders/d0/gnksqzwn2fn46fjgrkp6045c0000gn/T/ipykernel_49841/938463883.py in <dictcomp>(.0)
      2 
      3 revd = {
----> 4         k: max((e for e in v if pd.notna(e)), key=lambda x: float(clean_currency(x)) if isinstance(x, str) else x)
      5         for k,v in d.items()
      6        }

ValueError: max() arg is an empty sequence

Answer 1

The problem is you are filtering all the nans out, leaving you with an empty sequence in some case. But there is no need to filter nans, as far as I can tell. The reasonable response for the maximum of all nans is nan, which is already the behavior you'd get, so just use something like this:

from pprint import pprint

def clean_currency(x):
    """ If the value is a string, then remove currency symbol and delimiters
    otherwise, the value is numeric and can be converted
    """
    if isinstance(x, str):
       x = x.replace('$', '').replace(',', '')
    return float(x)



d = {
     'shores': ['$0.00'],
     'Broderick': ['$0.00', '$0.00', '$0.00'],
     '100 Broderick': ['$0.00', '$1,142,070.00', '$0.00', '$0.00', '$0.00'],
     '1001 Orange Grove': ['$0.00'],
     '1008 Hyde': [float('nan')]
    }


# Keep the highest value from the list.

revd = {
    k: max(v, key=clean_currency)
    for k,v in d.items()
}

pprint(revd, sort_dicts=False)

which outputs:

{'shores': '$0.00',
 'Broderick': '$0.00',
 '100 Broderick': '$1,142,070.00',
 '1001 Orange Grove': '$0.00',
 '1008 Hyde': nan}