np.where 和 np.select 返回错误。 AttributeError: 'numpy.ndarray' 对象没有属性 'select'

Question

Trying to use np.select or np.where to run a method on a DataFrame Column. Either way I do it, I get the error saying the "object has no attribute 'select' (or 'where' if I use np.where). When I convert the column and check the specific column type, it returns as an numpy.ndarray. According to the numpy docs, both np.where and np.select should work on numpy arrays.

Checking type on the Series I converted to a numpy array, after checking the type I try the method and receive the errors:

type(np.array(daily_std_df['Daily_Std']))
numpy.ndarray

type(np.array(x))
numpy.ndarray

### Apply method to the data frame using np.where OR np.select:
x = daily_std_df['Daily_Std']
conditionList = [x > SP_500_std, x < SP_500_std]
choiceList = ['High', "Not High"]
daily_std_df["Risk"] = np.array(x).select(conditionList, choiceList)
daily_std_df
---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
<ipython-input-51-42673f1eeafa> in <module>
      3 conditionList = [x > SP_500_std, x < SP_500_std]
      4 choiceList = ['High', "Not High"]
----> 5 daily_std_df["Risk"] = np.array(x).select(conditionList, choiceList)
      6 daily_std_df

AttributeError: 'numpy.ndarray' object has no attribute 'select'

Answer 1

It is not np.array which has the select attribute, it's just simply np that has the attribute.

So instead of:

daily_std_df["Risk"] = np.array(x).select(conditionList, choiceList)

Try this:

daily_std_df["Risk"] = np.select(conditionList, choiceList)

Read the docs for more examples.