Question

我在Python中有一个数据框，其中包含以下字段和类型：

dd = {'sp': ['1,1', '3,2'], 'tt': ['a', 'b'], 'zz': [4.5, 2.1]}

for k in dd:
    for i in range(len(dd[k])):
        if type(dd[k][i]) == str and ',' in dd[k][i]:
            dd[k][i] = float(dd[k][i].replace(',', '.'))

print(dd)

我希望浏览我的数据框并将偏移量（以秒为单位）添加到已发送的字段中。

我该怎么做？我尝试过以下方法：

sent         datetime64[ns]
tz_offset             int64

但是失败了，出现以下错误：

files['sent'] = files['sent'].apply(lambda x: x + np.timedelta64(files['tz_offset']), 's')

编辑：使用Pandas

Answer 1

矢量化方法是转换类型并添加df['sent'] + df['tz_offset'].astype("timedelta64[s]")

In [346]: df
Out[346]:
                 sent  tz_offset
0 2011-01-01 00:00:00          2
1 2011-01-01 00:01:00          0
2 2011-01-01 00:02:00          4
3 2011-01-01 00:03:00          0
4 2011-01-01 00:04:00          4
5 2011-01-01 00:05:00          4
6 2011-01-01 00:06:00          4
7 2011-01-01 00:07:00          1
8 2011-01-01 00:08:00          4
9 2011-01-01 00:09:00          4

In [347]: df['sent'] + df['tz_offset'].astype("timedelta64[s]")
Out[347]:
0   2011-01-01 00:00:02
1   2011-01-01 00:01:00
2   2011-01-01 00:02:04
3   2011-01-01 00:03:00
4   2011-01-01 00:04:04
5   2011-01-01 00:05:04
6   2011-01-01 00:06:04
7   2011-01-01 00:07:01
8   2011-01-01 00:08:04
9   2011-01-01 00:09:04
dtype: datetime64[ns]


In [348]: df.dtypes
Out[348]:
sent         datetime64[ns]
tz_offset             int32
dtype: object

或在apply

上使用np.timedelta64()

In [349]: df.apply(lambda x: x['sent'] + np.timedelta64(x['tz_offset'], 's'), axis=1)
Out[349]:
0   2011-01-01 00:00:02
1   2011-01-01 00:01:00
2   2011-01-01 00:02:04
3   2011-01-01 00:03:00
4   2011-01-01 00:04:04
5   2011-01-01 00:05:04
6   2011-01-01 00:06:04
7   2011-01-01 00:07:01
8   2011-01-01 00:08:04
9   2011-01-01 00:09:04
dtype: datetime64[ns]

或，使用pd.offsets.timedelta(seconds=)又名pd.Timedelta(seconds=)

In [350]: df.apply(lambda x: x['sent'] + pd.Timedelta(seconds=x['tz_offset']), axis=1)
Out[350]:
0   2011-01-01 00:00:02
1   2011-01-01 00:01:00
2   2011-01-01 00:02:04
3   2011-01-01 00:03:00
4   2011-01-01 00:04:04
5   2011-01-01 00:05:04
6   2011-01-01 00:06:04
7   2011-01-01 00:07:01
8   2011-01-01 00:08:04
9   2011-01-01 00:09:04
dtype: datetime64[ns]

如何在Python datetime64 [ns]字段中添加x秒？

1 个答案: