Reindex将值替换为NaN

Question

我正在尝试填充数据帧中缺少的时间。当我重新建立索引时，丢失的时间会被填充，但我的原始值会消失，所有内容都会变成一个NaN。我的代码如下：

data = {'date':["2010-03-01 00:03:00","2010-03-01 00:05:00"], 'num':["1","2"]}
df_test = pd.DataFrame(data,columns = ['date','num'])
pd.to_datetime(df_test['date'], format='%Y-%m-%d %H:%M:%S')
df_test.set_index('date', inplace=True)
df_test2 = df_test.reindex(pd.date_range(df_test.index[0], df_test.index[-1], freq='Min')) 

df_test

                       num
date    
2010-03-01 00:03:00     1
2010-03-01 00:05:00     2

df_test2

                        num
2010-03-01 00:03:00     NaN
2010-03-01 00:04:00     NaN
2010-03-01 00:05:00     NaN

Answer 1

这里有必要将转换后的列重新分配给datetimes，因为to_datetime不能就地工作：

data = {'date':["2010-03-01 00:03:00","2010-03-01 00:05:00"], 'num':["1","2"]}
df_test = pd.DataFrame(data,columns = ['date','num'])
df_test['date'] = pd.to_datetime(df_test['date'], format='%Y-%m-%d %H:%M:%S')
df_test.set_index('date', inplace=True)

df_test2 = df_test.reindex(pd.date_range(df_test.index[0], df_test.index[-1], freq='Min')) 

print (df_test2)

2010-03-01 00:03:00    1
2010-03-01 00:04:00  NaN
2010-03-01 00:05:00    2

另一种使用DataFrame.asfreq的解决方案

df_test2 = df_test.asfreq('Min')
print (df_test2)
                     num
date                    
2010-03-01 00:03:00    1
2010-03-01 00:04:00  NaN
2010-03-01 00:05:00    2