CODEWITHSUNDEEP
pythonpandassum
Neeraj Hanumante
Neeraj Hanumante

Reputation: 1684

Pandas sum includes column headers

Background

I have a dataset of about 200 countries (rows) for different time periods (columns). The Pandas dataframe of this dataset is as follows:

data = {'Country': ['Afghanistan', 'Albania', 'Algeria', 'Andorra', 'Angola'],
        '1958-1962': [0, 0, 0, 0, 0],
        '2008-2012': [0.0, 0.0, 8.425, 0.0, 0.0],
        '2013-2017': [0.0, 0.0, 10.46, 0.0, 0.0]}

df = pd.DataFrame(data)

     Country  1958-1962  2008-2012  2013-2017
 Afghanistan          0      0.000       0.00
     Albania          0      0.000       0.00
     Algeria          0      8.425      10.46
     Andorra          0      0.000       0.00
      Angola          0      0.000       0.00

I am trying to obtain sum of all the values in each column using the following code.

y_data = []

period_list = list(df)
period_list.remove('Country')

for x in period_list:
    y_data.append(df[x].sum())

Error

TypeError: unsupported operand type(s) for +: 'int' and 'str'

Process finished with exit code 1

For some reason, Pandas is also including the header in the sum process. How do I resolve this?

Other tests

I tested the sum function on the following dataframe using df.sum() , and it appropriately produced the sum of numbers for each column as 18, 20, 20, 19.

df = pd.DataFrame({"A":[5, 3, 6, 4], 
                   "B":[11, 2, 4, 3], 
                   "C":[4, 3, 8, 5], 
                   "D":[5, 4, 2, 8]})

The output of print(df.drop("Country",axis=1).dtypes) is as follows:

1958-1962    object
1963-1967    object
1968-1972    object
1973-1977    object
1978-1982    object
1983-1987    object
1988-1992    object
1993-1997    object
1998-2002    object
2003-2007    object
2008-2012    object
2013-2017    object
dtype: object

Solution

I used df = df.apply(pd.to_numeric, errors='ignore') to convert the objects into numbers and that resolved the issue.

Upvotes: 0

Views: 1112

Answers (1)

user12159041
user12159041

Reputation:

Convert the columns you want to sum from objects to numeric and then drop Country column before making sum in the rest of columns.

Refer this link for converting from object to numeric

Upvotes: 1

Related Questions