Looping through dataframe with conditions-CodePudding

This is my input data, which is stored in dataframe df.

Now i want to change all the values in column B to Yearly format. Here is my code:

D = []
for i in df['B']:
    for j in df['C']:
        if j == 'Year':
            D.append(int(i)/1) 
        elif j == 'Month':
            D.append(int(i)/12)
        elif j == 'Day':
            D.append(int(i)/365)

print(len(df))
print(len(D))

While my original df only has len of 10, the output (list D) has len of 100. Anyone knows how to fix the issue here?

CodePudding user response：

You can try with map

df['D'] = df['B'].div(df['C'].map({'Year':1, 'Month':12, 'Day':365})

CodePudding user response：

B is of size 10 and C is of size 10. You loop through C 10 times. This will generate D of size 100. Because they are of the same size, you only need 1 for loop to populate D.

CodePudding user response：

Your code iterates 10 X 10 times because for each i of df['B'] it iterates throughout all the rows of df['C'].

https://pandas.pydata.org/docs/user_guide/basics.html#iteration

D = []
for row in df.itertuples():
   if row.C == 'Year':
      D.append(int(i)/1) 
   elif row.C == 'Month':
      D.append(int(i)/12)
   elif row.C == 'Day':
      D.append(int(i)/365)

CodePudding user response：

You should aplly a function using the df.apply() method.

Just define your logic inside a function to return the value you want for each row, like:

def convertValues(row):
    if row['C'] == 'Year':
        return int(row['B'])
    elif row['C'] == 'Month':
        return int(row['B'])/12
    elif row['C'] == 'Day':
        return int(row['C'])/365
    return 'Invalid String'

Then you simply apply to the dataframe:

yearly_values = df.apply(convertValues, axis=1)

Your result will be a Pandas Series, which you can then cast to list or do whatever you want to do.

You can also create a new column on that dataframe with the respective values using:

df['D'] = df.apply(convertValues, axis=1)