CODEWITHSUNDEEP
pythonpandas
Pebeto
Pebeto

Reputation: 140

value_counts return counts for the same value?

Here is the following class

class MyClass: 

    def __init__(self, cat): 
        if cat not in ['STR', 'MTR', 'CA', 'Other', 'None']:
            raise ValueError
        self.cat = cat
        
    def __add__(self, other): 
        if self.cat == 'STR' or other.cat == 'STR':
            return MyClass('STR')
        if self.cat == 'MTR' or self.cat == 'CA' or other.cat == 'MTR' or other.cat == 'CA':
            return MyClass('MTR')
        return MyClass('None')
    
    def __eq__(self, other): 
        if isinstance(other, str):
            return self.cat == other
        elif isinstance(other, MyClass): 
            return self.cat == other.cat
        else:
            raise ValueError
            
    def __ne__(self, other): 
        if isinstance(other, str):
            return self.cat != other
        elif isinstance(other, MyClass): 
            return self.cat != other.cat
        else:
            raise ValueError    
            
    def __str__(self): 
        return self.cat

I am trying to add arrays of MyClass instances indexed by dates:

start_date =  pd.Timestamp(datetime.date(2017, 1, 1))
end_date =  pd.Timestamp(datetime.date(2017, 1, 31))
ref = pd.Series(MyClass('None'), index=pd.date_range(start_date, end_date))
ref.value_counts()

None    31
dtype: int64

start_date =  pd.Timestamp(datetime.date(2017, 1, 10))
end_date =  pd.Timestamp(datetime.date(2017, 1, 20))
test = pd.Series(MyClass('STR'), index=pd.date_range(start_date, end_date))
test.value_counts()

STR    11
dtype: int64

ref = ref.add(test, fill_value=MyClass('None'))
ref.value_counts()

I do not understand the results from value_counts above

None    19
STR     11
None     1
dtype: int64

Why None appears twice?

Thank you

Upvotes: 0

Views: 100

Answers (1)

vamsi_s
vamsi_s

Reputation: 201

i think hash value of your object is required for pandas to check uniqueness.

can you add below to your class -

def __hash__(self):
    return self.cat.__hash__()

after adding this, you should get the below value -

None    20
STR     11
dtype: int64

Upvotes: 1

Related Questions