Pandas.DataFrame 数据处理

df = pd.DataFrame()
df['id'] = [1,1,2]
df['movie'] = ['2012','The Pursuit of Happyness',  'Avatar']

id = [1,1,2]
movie = ['2012','The Pursuit of Happyness',  'Avatar']
df = pd.DataFrame({'id': id, 'movie': movie})

df.columns = ['id', 'movie', 'score', 'time']

df.rename(columns={0: 'id',1:'movie', 2:'score', 3:'time'}, inplace= True)

df.loc[0:2, ['movie','score']]

df.iloc[0:2, 1:3]

def is_favorite(x) :
    if x < 3 :
        return "No"
    else :
        return "Yes" 

df['favorite'] = df['score'].map(is_favorite)

swap = {"Yes": 1, "No": 0}
df.replace(swap, inplace=True)

t = pd.DatetimeIndex(df['time'])
df['year'] = t.year
df['month'] = t.month
df['day'] = t.day
df['dayofweek'] = t.dayofweek

df['score'].value_counts()

import matplotlib.pyplot as plt
df['score'].value_counts().plot.bar(figsize=(15,10))
plt.show()

df['score'].isnull().mean()

df['score'].fillna(60, inplace = True)

df = pd.concat([df1, df2], axis=0,ignore_index=True)

df = pd.merge(df, df_attr, how='left', on='id')

df.loc[df['is_favorite'] == 0]

ml = ['2012', 'Avatar']
df.loc[df['movie'].isin(ml)]

df.drop_duplicates(subset=['id','movie'], keep='first')