使用numpy.where向数据框输入值

Question

I'm trying to entering values to dataframe based on a condition. I have movie attribute columns like 'AttributeHorror' or 'AttributeRomance' etc. Columns are filled with NaN. If the condition is met, it have to replace with 1. if not, 0.

I tried the following code but it didnt return correct.

film_data['AttributeOlumsuz'] 
       =np.where(film_data['FilmAttributes1']=='Olumsuz', '1', '0')


film_data['AttributeSiddet'] = np.where(film_data['FilmAttributes1']=='Siddet', '1', '0')


film_data['AttributeAltyazi'] = 
np.where(film_data['FilmAttributes1']=='Altyazili', '1', '0')

film_data['AttributeGerilim'] = 
np.where(film_data['FilmAttributes1']=='Gerilim', '1', '0')

This code fills every cell incorrect. Where am i making mistake?

Answer 1

A simple list comprehension should work.

There are two parts:

1) iterate over the original values

2) list comprehension conditional assignment

These are commented in the code below.

film_data['AttributeOlumsuz'] = [
    '1' if v=='olumsuz' else '0' #2 list comprehension conditional assignment here
    for v in film_data['FilmAttributes1'] #1 iterate over original values here
]

Answer 2

If numpy.where() does not work.

Assuming you are reading your data using pandas.DataFrame, you can try:

First Possible Solution:

df.loc[df['AttributeOlumsuz'] == 'Olumsuz', 'Tagger_Column'] = '1'

Second Possible Solution:

You can also use .map() method.

your_dict = {'Olumsuz': '1'}
df['Tagger_Column'] = df['AttributeOlumsuz'].map(your_dict)

Result of map for False entries is NaN , if you don't like that you can use fillna with parameter equals to whatever value you please:

df['Tagger_Column'].fillna(value = '0') # to replicate your condition in numpy.where() above