如何在python数据帧中仅选择某个标签后的3行？

我有这样的数据：

Month   Time          Label
Apr     19:10:11      Query
Apr     19:10:11      Not Command
Apr     19:10:12      Not Command
Apr     19:10:12      Sending
Apr     19:10:13      Not Command
Apr     19:10:13      Not Command
Apr     19:10:14      Not Command
Apr     19:10:14      Not Command
Apr     19:10:14      Answer
Apr     19:10:14      Not Command
Apr     19:10:15      Not Command
Apr     19:10:15      Not Command

我想将表放入每个标签中，这些标签不等于

not Command

，比如

Query

、

Sending

和

Answer

。我想选择标签后始终有3行。这就是我想要的结果

Month   Time          Label
Apr     19:10:11      Query
Apr     19:10:11      Not Command
Apr     19:10:12      Not Command
Apr     19:10:12      Sending
Apr     19:10:13      Not Command
Apr     19:10:13      Not Command
Apr     19:10:14      Answer
Apr     19:10:14      Not Command
Apr     19:10:15      Not Command

---

可能吗？我有一个想法，使新的标签，迭代数字时，遇到标签，但我不知道如何。所以也许有更好的方法来解决这个问题。感谢您

如果您认为选择

3

或更少的行创建掩码依据，并使用以下内容进行筛选：

如果需要，请选择3行（如果存在3行或更多行），否则不要选择添加新掩码，并按位过滤

和

-

&

：

print (df)
   Month      Time        Label
0    Apr  19:10:11        Query
1    Apr  19:10:11  Not Command <- only 2 rows - omited
2    Apr  19:10:12      Sending
3    Apr  19:10:13  Not Command
4    Apr  19:10:13  Not Command
5    Apr  19:10:14  Not Command
6    Apr  19:10:14  Not Command
7    Apr  19:10:14       Answer
8    Apr  19:10:14  Not Command
9    Apr  19:10:15  Not Command
10   Apr  19:10:15  Not Command

s = df['Label'].ne('Not Command').cumsum()

df = df[s.map(s.value_counts()).ge(3) & s.groupby(s).cumcount().lt(3)]
print (df)
  Month      Time        Label
2   Apr  19:10:12      Sending
3   Apr  19:10:13  Not Command
4   Apr  19:10:13  Not Command
7   Apr  19:10:14       Answer
8   Apr  19:10:14  Not Command
9   Apr  19:10:15  Not Command

打印（df）
月份时间标签
4月0日19:10:11查询
4月1日19:10:11不是命令是的，有可能。但让我们看看你到目前为止都做了些什么。
print (df)
   Month      Time        Label
0    Apr  19:10:11        Query
1    Apr  19:10:11  Not Command <- only 2 rows - omited
2    Apr  19:10:12      Sending
3    Apr  19:10:13  Not Command
4    Apr  19:10:13  Not Command
5    Apr  19:10:14  Not Command
6    Apr  19:10:14  Not Command
7    Apr  19:10:14       Answer
8    Apr  19:10:14  Not Command
9    Apr  19:10:15  Not Command
10   Apr  19:10:15  Not Command

s = df['Label'].ne('Not Command').cumsum()

df = df[s.map(s.value_counts()).ge(3) & s.groupby(s).cumcount().lt(3)]
print (df)
  Month      Time        Label
2   Apr  19:10:12      Sending
3   Apr  19:10:13  Not Command
4   Apr  19:10:13  Not Command
7   Apr  19:10:14       Answer
8   Apr  19:10:14  Not Command
9   Apr  19:10:15  Not Command