np.where: "SyntaxError: keyword can't be an expression"

Question

I am trying to incorporate the following logic

In Column A, where column B = Stack then Column A * 100 otherwise keep column A as-is.

df['Value'] = np.where(df['columnB'] = 'Stack', df['Value'] * 100)

why am I getting a SyntaxError: keyword can't be an expression here?

Answer 1

You must supply a condition and two outcomes for the binary condition

Without knowing much about you dataframe I believe you should be doing something like the following:

df['Value'] = np.where(df['columnB'] == 'Stack', df['columnA']*100, df['Value'])

This is because in the documentation it states:

numpy.where(condition[, x, y])
    Return elements chosen from x or y depending on condition.

Parameters: 
    condition : array_like, bool
    Where True, yield x, otherwise yield y.

So df['Value'] would be populated with columnA times 100 if columnB is 'Stack', otherwise, it keeps the value that is stored at Value