按值差异（时间戳）将一列分成 N 组

Question

.csv 格式的样本数据

| No.|   IP     |      Unix_time     |    # integer unix time
| 1  | 1.1.1.1  |     1563552000     |    # equivalent to 12:00:00 AM
| 2  | 1.1.1.1  |     1563552030     |    # equivalent to 12:00:30 AM
| 3  | 1.1.1.1  |     1563552100     |    # equivalent to 12:01:40 AM
| 4  | 1.1.1.1  |     1563552110     |    # equivalent to 12:01:50 AM
| 5  | 1.1.1.1  |     1563552180     |    # equivalent to 12:03:00 AM
| 6  | 1.2.3.10 |     1563552120     |    

这是使用 pandas groupby( ) 和 get_group( ) 函数的当前工作代码：

data = pd.read_csv(some_path, header=0)
root = data.groupby('IP')

for a in root.groups.keys():
    t = root.get_group(a)['Unix_time']
    print(a + 'has' + t.count() + 'record')

您将看到以下结果：

1.1.1.1 has 5 record
1.2.3.10 has 1 record

现在，我想根据上面的代码进行一些改进。

对于相同的 IP 值（例如，1.1.1.1），我想根据最大时间间隔（例如，60 秒）制作更多 子组 ，并计算每个子组中的元素数量。例如，在上面的示例数据中：

从第 1 行开始：第 2 行 Unix_time 值在 60​​ 秒内，但第 3 行超过 60 秒。

因此，第 1-2 行是一个组，第 3-4 行是一个单独的组，第 5 行是一个单独的组。换句话说，组 '1.1.1.1' 现在有 3 个子组。结果应该是：

1.1.1.1 start time 1563552000 has 2 record within 60 secs
1.1.1.1 start time 1563552100 has 2 record within 60 secs
1.1.1.1 start time 1563552150 has 1 record within 60 secs
1.2.3.10 start time 1563552120 has 1 record within 60 secs

如何制作？

Answer 1

你可以使用pd.Grouper:

df['datetime'] = pd.to_datetime(df['Unix_time'], unit='s')
for n, g in df.groupby(['IP', pd.Grouper(freq='60s', key='datetime')]):
    print(f'{n[0]} start time {g.iloc[0, g.columns.get_loc("Unix_time")]} has {len(g)} records within 60 secs')

输出：

1.1.1.1  start time 1563552000 has 2 records within 60 secs
1.1.1.1  start time 1563552100 has 2 records within 60 secs
1.1.1.1  start time 1563552150 has 1 records within 60 secs
1.2.3.10 start time 1563552120 has 1 records within 60 secs

---

使用“根”和整数：

root = df.groupby(['IP',df['Unix_time']//60])

for n, g in root:
     print(f'{n[0]} start time {g.iloc[0, g.columns.get_loc("Unix_time")]} has {len(g)} records within 60 secs')

输出：

1.1.1.1  start time 1563552000 has 2 records within 60 secs
1.1.1.1  start time 1563552100 has 2 records within 60 secs
1.1.1.1  start time 1563552150 has 1 records within 60 secs
1.2.3.10 start time 1563552120 has 1 records within 60 secs