TypeError: unsupported operand type(s) for &: 'tuple' and 'tuple'-CodePudding

I'm trying to do some data validation with numpy and I don't really understand this error. see code below:

conditions = [
    (np.where(df['DD'].eq('No DD Required'))),
    (np.where(df['DD'].eq('DD Required'))) & (np.where(df['On Direct Debit'].eq('No'))),
    (np.where(df['DD'].eq('DD Required'))) & (np.where(df['On Direct Debit'].eq('Yes')))
    ]


values = ['Pass', 'Fail', 'Pass']

df['DDValidation'] = np.select(conditions, values)

CodePudding user response：

Your & operator is not the one you wanted. & in python is a bitwise AND, not the logical AND, which is and.

Change your code to something like

conditions = [
    (np.where(df['DD'].eq('No DD Required'))),
    (np.where(df['DD'].eq('DD Required'))) and (np.where(df['On Direct Debit'].eq('No'))),
    (np.where(df['DD'].eq('DD Required'))) and (np.where(df['On Direct Debit'].eq('Yes')))
    ]

for more info, a discussion of the distinction between these different ANDs

CodePudding user response：

This has more of a chance of working. You haven't provided a sample df, so I can't test it.

conditions = [
    df['DD'].eq('No DD Required',
    (df['DD'].eq('DD Required')) & (df['On Direct Debit'].eq('No')),
    (df['DD'].eq('DD Required')) & (df['On Direct Debit'].eq('Yes'))
    ]

Look at what np.select expects as its first argument:

condlist : list of bool ndarrays

Each of those lines needs to a bool numpy array.

When developing, and debugging, code, test each step. Don't guess or assume. Look at what np.where(...) produces. Does that look at all like a bool array? And read the docs - for np.where and np.select for a start.