Python pandas 通过 dt 访问器有效地将日期时间转换为时间戳

我觉得你需要先转换成 numpy array by values 并转换为 int64 - 输出在 ns，所以需要除以10 ** 9:

df['ts'] = df.datetime.values.astype(np.int64)//10 ** 9打印 (df)日期时间 ts0 2016-01-01 00:00:01 14516064011 2016-01-01 01:00:01 14516100012 2016-01-01 02:00:01 14516136013 2016-01-01 03:00:01 14516172014 2016-01-01 04:00:01 14516208015 2016-01-01 05:00:01 14516244016 2016-01-01 06:00:01 14516280017 2016-01-01 07:00:01 14516316018 2016-01-01 08:00:01 14516352019 2016-01-01 09:00:01 145163880110 2016-01-01 10:00:01 145164240111 2016-01-01 11:00:01 145164600112 2016-01-01 12:00:01 145164960113 2016-01-01 13:00:01 145165320114 2016-01-01 14:00:01 145165680115 2016-01-01 15:00:01 145166040116 2016-01-01 16:00:01 145166400117 2016-01-01 17:00:01 145166760118 2016-01-01 18:00:01 145167120119 2016-01-01 19:00:01 145167480120 2016-01-01 20:00:01 145167840121 2016-01-01 21:00:01 145168200122 2016-01-01 22:00:01 145168560123 2016-01-01 23:00:01 145168920124 2016-01-02 00:00:01 1451692801

to_timestamp 用于将 从周期索引转换为日期时间索引.

I have a dataframe with some (hundreds of) million of rows. And I want to convert datetime to timestamp effectively. How can I do it?

My sample df:

df = pd.DataFrame(index=pd.DatetimeIndex(start=dt.datetime(2016,1,1,0,0,1),
    end=dt.datetime(2016,1,2,0,0,1), freq='H'))
    .reset_index().rename(columns={'index':'datetime'})

df.head()

             datetime
0 2016-01-01 00:00:01
1 2016-01-01 01:00:01
2 2016-01-01 02:00:01
3 2016-01-01 03:00:01
4 2016-01-01 04:00:01

Now I convert datetime to timestamp value-by-value with .apply() but it takes a very long time (some hours) if I have some (hundreds of) million rows:

df['ts'] = df[['datetime']].apply(lambda x: x[0].timestamp(), axis=1).astype(int)

df.head()

             datetime          ts
0 2016-01-01 00:00:01  1451602801
1 2016-01-01 01:00:01  1451606401
2 2016-01-01 02:00:01  1451610001
3 2016-01-01 03:00:01  1451613601
4 2016-01-01 04:00:01  1451617201

The above result is what I want.

If I try to use the .dt accessor of pandas.Series then I get error message:

df['ts'] = df['datetime'].dt.timestamp

AttributeError: 'DatetimeProperties' object has no attribute 'timestamp'

If I try to create eg. the date parts of datetimes with the .dt accessor then it is much more faster then using .apply():

df['date'] = df['datetime'].dt.date

df.head()

             datetime          ts        date
0 2016-01-01 00:00:01  1451602801  2016-01-01
1 2016-01-01 01:00:01  1451606401  2016-01-01
2 2016-01-01 02:00:01  1451610001  2016-01-01
3 2016-01-01 03:00:01  1451613601  2016-01-01
4 2016-01-01 04:00:01  1451617201  2016-01-01

I want something similar with timestamps...

But I don't really understand the official documentation: it talks about "Converting to Timestamps" but I don't see any timestamps there; it just talks about converting to datetime with pd.to_datetime() but not to timestamp...

pandas.Timestamp constructor also doesn't work (returns with the below error):

df['ts2'] = pd.Timestamp(df['datetime'])

TypeError: Cannot convert input to Timestamp

pandas.Series.to_timestamp also makes something totally different that I want:

df['ts3'] = df['datetime'].to_timestamp

df.head()

             datetime          ts                                                ts3
0 2016-01-01 00:00:01  1451602801  <bound method Series.to_timestamp of 0    2016...
1 2016-01-01 01:00:01  1451606401  <bound method Series.to_timestamp of 0    2016...
2 2016-01-01 02:00:01  1451610001  <bound method Series.to_timestamp of 0    2016...
3 2016-01-01 03:00:01  1451613601  <bound method Series.to_timestamp of 0    2016...
4 2016-01-01 04:00:01  1451617201  <bound method Series.to_timestamp of 0    2016...

Thank you!!

I think you need convert first to numpy array by values and cast to int64 - output is in ns, so need divide by 10 ** 9:

df['ts'] = df.datetime.values.astype(np.int64) // 10 ** 9
print (df)
              datetime          ts
0  2016-01-01 00:00:01  1451606401
1  2016-01-01 01:00:01  1451610001
2  2016-01-01 02:00:01  1451613601
3  2016-01-01 03:00:01  1451617201
4  2016-01-01 04:00:01  1451620801
5  2016-01-01 05:00:01  1451624401
6  2016-01-01 06:00:01  1451628001
7  2016-01-01 07:00:01  1451631601
8  2016-01-01 08:00:01  1451635201
9  2016-01-01 09:00:01  1451638801
10 2016-01-01 10:00:01  1451642401
11 2016-01-01 11:00:01  1451646001
12 2016-01-01 12:00:01  1451649601
13 2016-01-01 13:00:01  1451653201
14 2016-01-01 14:00:01  1451656801
15 2016-01-01 15:00:01  1451660401
16 2016-01-01 16:00:01  1451664001
17 2016-01-01 17:00:01  1451667601
18 2016-01-01 18:00:01  1451671201
19 2016-01-01 19:00:01  1451674801
20 2016-01-01 20:00:01  1451678401
21 2016-01-01 21:00:01  1451682001
22 2016-01-01 22:00:01  1451685601
23 2016-01-01 23:00:01  1451689201
24 2016-01-02 00:00:01  1451692801

to_timestamp is used for converting from period to datetime index.