Python Pandas阅读时间

Question

When I am reading a time data from an xlsx file into pandas. It reads as a decimal value

Example :9:23:27 AM is read as .391284722 I can fix it by converting it into time using format cell and select time. But I would prefer to use pandas all the way through and not excel.

When I call the value and convert it into a date time object df.TIME=pd.to_datetime(df.TIME)

It changes to this date 1970-01-01 Desired time is 9:23:27 AM

Any help is greatly appreciated. Thank you

Answer 1

Demo:

read that column as string:

df = pd.read_excel(filename, dtype={'col_name':str})

In [51]: df
Out[51]:
          time
0   9:23:27 AM
1  12:59:59 AM

In [52]: df['time2'] = pd.to_timedelta(df['time'])

In [53]: df
Out[53]:
          time    time2
0   9:23:27 AM 09:23:27
1  12:59:59 AM 12:59:59

In [54]: df.dtypes
Out[54]:
time              object
time2    timedelta64[ns]
dtype: object

UPDATE: in order to convert a float number (# of seconds) read from Excel try the following:

Source DF:

In [85]: df
Out[85]:
       time
0  0.391285
1  0.391285
2  0.391285

Solution:

In [94]: df['time2'] = pd.to_timedelta((df['time'] * 86400).round(), unit='s')

In [95]: df
Out[95]:
       time    time2
0  0.391285 09:23:27
1  0.391285 09:23:27
2  0.391285 09:23:27

In [96]: df.dtypes
Out[96]:
time             float64
time2    timedelta64[ns]
dtype: object

Answer 2

The question could use some clarifying for an end-purpose for the time-column. For general purposes though, try using the format keyword in to_datetime .

df.TIME=pd.to_datetime(df.TIME, format='%I:%M%S %p')

See this website for formatting: http://strftime.org/