Add new column to Python Pandas DataFrame based on multiple conditions

Question

I have a dataset with various columns as below:

discount tax total subtotal productid 3.98 1.06 21.06 20 3232 3.98 1.06 21.06 20 3232 3.98 6 106 100 3498 3.98 6 106 100 3743 3.98 6 106 100 3350 3.98 6 106 100 3370 46.49 3.36 66.84 63 695

Now, I need to add a new column Class and assign it the value of 0 or 1 on the base of the following conditions:

if:
    discount > 20%
    no tax
    total > 100
then the Class will 1
otherwise it should be 0

I have done it with a single condition but I don't how can I accomplish it under multiple conditions.

Here's wIat i have tried:

df_full['Class'] = df_full['amount'].map(lambda x: 1 if x > 100 else 0)

I have taken a look at all other similar questions but couldn't find any solution for my problem.I have tried all of the above-mentioned posts but stuck on this error:

TypeError: '>' not supported between instances of 'str' and 'int'

Here's in the case of first posted answer, i have tried it as:

df_full['class'] = np.where( ( (df_full['discount'] > 20) & (df_full['tax'] == 0 ) & (df_full['total'] > 100) & df_full['productdiscount'] ) , 1, 0)

Answer 1

You can apply an arbitrary function across a dataframe row using DataFrame.apply.

In your case, you could define a function like:

def conditions(s):
    if (s['discount'] > 20) or (s['tax'] == 0) or (s['total'] > 100):
        return 1
    else:
        return 0

And use it to add a new column to your data:

df_full['Class'] = df_full.apply(conditions, axis=1)

Answer 2

Judging by the image of your data is rather unclear what you mean by a discount 20%.

However, you can likely do something like this.

df['class'] = 0 # add a class column with 0 as default value

# find all rows that fulfills your conditions and set class to 1
df.loc[(df['discount'] / df['total'] > .2) & # if discount is more than .2 of total 
       (df['tax'] == 0) & # if tax is 0
       (df['total'] > 100), # if total is > 100 
       'class'] = 1 # then set class to 1

Note that & means and here, if you want or instead use |.