Python Dataframe - Use column names as index, and sum of columns as values-CodePudding

I have a dataframe where the column names are dates, and values are number of cases. However, there are 300 daily columns. How do I create a series where the index is the column names (dates), and values are the sum of number of days for each column?

Input:

      Country     1/1/2020   1/2/2020  ...  12/31/2020
  0   Afganistan  50         100       ...  500
  1   Albania     20         50        ...  50
  ...
  99  Zimbabwe   6          10         ...  5

Desired output (pd.Series):

1/1/2020    76
1/2/2020    160
...
12/31/2020  555

CodePudding user response：

Use DataFrame.sum with numeric_only=True parameter:

s = df.sum(numeric_only=True)
print (s)
1/1/2020       76
1/2/2020      160
12/31/2020    555
dtype: int64

CodePudding user response：

You can set_index to "Country" and sum:

out = df.set_index('Country').sum()

or drop "Country" and sum:

out = df.drop(columns='Country').sum()

Output:

1/1/2020       76
1/2/2020      160
12/31/2020    555
dtype: int64

CodePudding user response：

Transposing and summing should do the trick:

df.T.sum()

Where df is your DataFrame