如何在python中找到事件的开始时间和结束时间？

Question

I have a data frame consists of column 1 ie event and column 2 is Datetime:

Sample data

 Event   Time
    0   2020-02-12 11:00:00
    0   2020-02-12 11:30:00
    2   2020-02-12 12:00:00
    1   2020-02-12 12:30:00
    0   2020-02-12 13:00:00
    0   2020-02-12 13:30:00
    0   2020-02-12 14:00:00
    1   2020-02-12 14:30:00
    0   2020-02-12 15:00:00
    0   2020-02-12 15:30:00

And I want to find start time and end time of each event:

Desired Data

 Event  EventStartTime  EventEndTime
    0   2020-02-12 11:00:00 2020-02-12 12:00:00
    2   2020-02-12 12:00:00 2020-02-12 12:30:00
    1   2020-02-12 12:30:00 2020-02-12 13:00:00
    0   2020-02-12 13:00:00 2020-02-12 14:30:00
    1   2020-02-12 14:30:00 2020-02-12 15:00:00

Note: EventEndTime is time when the event changes the value say from value 1 to got change to 0 or any other value or vice versa

Answer 1

Here is a method that can get the results without a for loop. I assume that the input data is read into a dataframe called df:

# Initialize the output df
dfout = pd.DataFrame()
dfout['Event'] = df['Event']
dfout['EventStartTime'] = df['Time']

Now, I create a variable called 'change' that tells you whether the event changed.

dfout['change'] = df['Event'].diff()

This is how dfout looks now:

   Event       EventStartTime  change
0      0  2020-02-12 11:00:00     NaN
1      0  2020-02-12 11:30:00     0.0
2      2  2020-02-12 12:00:00     2.0
3      1  2020-02-12 12:30:00    -1.0
4      0  2020-02-12 13:00:00    -1.0
5      0  2020-02-12 13:30:00     0.0
6      0  2020-02-12 14:00:00     0.0
7      1  2020-02-12 14:30:00     1.0
8      0  2020-02-12 15:00:00    -1.0
9      0  2020-02-12 15:30:00     0.0

Now, I go on to remove the rows where the event did not change:

dfout = dfout.loc[dfout['change'] !=0 ,:]

This will now leave me with rows where the event has changed.

Next, the event end time of the current event is the start time of the next event.

dfout['EventEndTime'] = dfout['EventStartTime'].shift(-1)

The dataframe looks like this:

   Event       EventStartTime  change         EventEndTime
0      0  2020-02-12 11:00:00     NaN  2020-02-12 12:00:00
2      2  2020-02-12 12:00:00     2.0  2020-02-12 12:30:00
3      1  2020-02-12 12:30:00    -1.0  2020-02-12 13:00:00
4      0  2020-02-12 13:00:00    -1.0  2020-02-12 14:30:00
7      1  2020-02-12 14:30:00     1.0  2020-02-12 15:00:00
8      0  2020-02-12 15:00:00    -1.0                  NaN

You may chose to remove the 'change' column and also the last row if not needed.

Answer 2

Assuming the dataframe is data :

current_event = None
result = []
for event, time in zip(data['Event'], data['Time']):
    if event != current_event:
        if current_event is not None:
            result.append([current_event, start_time, time])
        current_event, start_time = event, time
data = pandas.DataFrame(result, columns=['Event','EventStartTime','EventEndTime'])

The trick is to save your event number; if the next event number is not the same as the saved one, the saved one has to be ended and a new one started.

Answer 3

Use group by and agg to get the output in desired format.

df =pd.DataFrame([['0',11],['1',12],['1',13],['0',15],['1',16],['3',11]],columns=['Event','Time'] )
df.groupby(['Event']).agg(['first','last']).rename(columns={'first':'start-event','last':'end-event'})

Output:

Event start-event   end-event   
0      11           15
1      12           16
3      11           11