Question

我想基于for-loop中的另一列data使用df查找连续的时间段，即时间段（使用开始和结束时间戳记定义），其中data> 20。在df中，timestamp被用作索引。我认为问题在于，在循环中，我没有正确指定从数据帧的索引列中选择行。

for-loop：

for i in range(len(df3)): 
    if i >0:

        activities = []          
        start_time = None          

        if (df.loc[i, 'data'] >= 20):                                   

            if start_time == None:   
                start_time = df.loc[i, 'timestamp']
        else:

            if start_time != None:
                end_time = df.loc[i-1, 'timestamp']

                duration = (end_time - start_time).seconds
                activities.append((duration, start_time, end_time))
                start_time = None 

return activities

df：

                        id      timestamp               data    Date        sig     events
timestamp                           
2020-01-15 06:12:49.213 40250   2020-01-15 06:12:49.213 20.0    2020-01-15  -1.0    1.0
2020-01-15 06:12:49.313 40251   2020-01-15 06:12:49.313 19.5    2020-01-15  1.0     0.0
2020-01-15 08:05:10.083 40256   2020-01-15 08:05:10.083 20.0    2020-01-15  1.0     0.0

返回：

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-11-9853026603d5> in <module>()
      9 
     10 
---> 11         if (df.loc[i, 'data'] >= 20):                                   
     12 
     13             if start_time == None:

7 frames
/usr/local/lib/python3.6/dist-packages/pandas/core/indexes/base.py in _invalid_indexer(self, form, key)
   3074         """
   3075         raise TypeError(
-> 3076             f"cannot do {form} indexing on {type(self)} with these "
   3077             f"indexers [{key}] of {type(key)}"
   3078         )

TypeError: cannot do index indexing on <class 'pandas.core.indexes.datetimes.DatetimeIndex'> with these indexers [1] of <class 'int'>

更新：

按照@jcaliz的建议，我尝试了以下代码，并针对不同的变化更改了return的缩进：

for i in range(len(df)): 
    if i >0:

        activities = []           
        start_time = None          


        if (df.iloc[I].data >= 20):                                   
            if start_time == None:   
                start_time = df.iloc[i].timestamp
        else:

            if start_time != None:
                end_time = df.iloc[i-1].timestamp

                duration = (end_time - start_time).seconds
                activities.append((duration, start_time, end_time))
                start_time = None


return activities

但有相同的错误：

  File "<ipython-input-24-d78e4605aebe>", line 31
    return activities
                            ^
SyntaxError: 'return' outside function

Answer 1

loc用于文本，而不是基于整数的索引，而应使用iloc。更改：

if (df.loc[i, 'data'] >= 20):

收件人

if (df.iloc[i].data >= 20):

对于其他loc之类的df.loc[i, 'timestamp']

也是如此

编辑：

更好的方法是不使用for循环

start_time与timestamp
end_time是前一个的timestamp
duration是秒数的差异

该过程将是：

# Assign previous record's timestamp as end time
df['end_time'] = df['timestamp'].shift(1)

df['duration'] = df.apply(lambda x: (x['end_time'] -
                                     x['timestamp']).seconds, axis=1)

如何从熊猫的索引列中选择行

1 个答案: