Python 如何标记值​；根据特定规则在一系列中？

我想在我的职业生涯中发现积极和消极的波动。那么如何标记数据呢
示例
我的数据：

我想要的结果是：

| date                | value | sign |
|---------------------|-------|------|
| 2018-09-06 00:00:03 | 0     | 1    |
| 2018-09-06 00:00:04 | 0     | 1    |
| 2018-09-06 00:00:05 | 1     | 1    |
| 2018-09-06 00:00:06 | 1     | 1    |
| 2018-09-06 00:00:07 | 2     | 1    |
| 2018-09-06 00:00:08 | -1    | 2    |
| 2018-09-06 00:00:09 | -5    | 2    |
| 2018-09-06 00:00:10 | 0     | 2    |
| 2018-09-06 00:00:11 | -6    | 2    |
| 2018-09-06 00:00:12 | 2     | 3    |
| 2018-09-06 00:00:13 | 0     | 3    |
| 2018-09-06 00:00:14 | 4     | 3    |

然后：

mydata.groupby（['sign']）.transform（'sum'））

使用+：

---

采样数据不包括正波和负波由零分隔的情况，例如

1 0 0-1

。以下是一个解决方案，涵盖了这一情况：

# mask the zeros
s = mydata['value'].eq(0)

# merge the zeros to the wave after them
m = np.sign(mydata['value']).mask(s).bfill()

# result
mydata['sign'] = m.diff().ne(0).cumsum()

---

如果您的数据有

100-1

，会发生什么情况？您好@ansev。谢谢你的回复。我查过你的密码了。但是输出有错误。如果你想检查结果：你好，广黄。谢谢你解决了我的问题。成功了！

mydata['sign']=(mydata['value']*mydata['value'].shift()<0).cumsum()+1
print(mydata)

                   date  value  sign
0   2018-09-06-00:00:03      0     1
1   2018-09-06-00:00:04      0     1
2   2018-09-06-00:00:05      1     1
3   2018-09-06-00:00:06      1     1
4   2018-09-06-00:00:07      2     1
5   2018-09-06-00:00:08     -1     2
6   2018-09-06-00:00:09     -5     2
7   2018-09-06-00:00:10      0     2
8   2018-09-06-00:00:11     -6     2
9   2018-09-06-00:00:12      2     3
10  2018-09-06-00:00:13      0     3
11  2018-09-06-00:00:14      4     3

mydata[['sum_value','sum_sign']]=mydata.groupby(['sign'])[['value','sign']].transform('sum')
print(mydata)
                   date  value  sign  sum_value  sum_sign
0   2018-09-06-00:00:03      0     1          4         5
1   2018-09-06-00:00:04      0     1          4         5
2   2018-09-06-00:00:05      1     1          4         5
3   2018-09-06-00:00:06      1     1          4         5
4   2018-09-06-00:00:07      2     1          4         5
5   2018-09-06-00:00:08     -1     2        -12         8
6   2018-09-06-00:00:09     -5     2        -12         8
7   2018-09-06-00:00:10      0     2        -12         8
8   2018-09-06-00:00:11     -6     2        -12         8
9   2018-09-06-00:00:12      2     3          2         6
10  2018-09-06-00:00:13      0     3          2         6

# mask the zeros
s = mydata['value'].eq(0)

# merge the zeros to the wave after them
m = np.sign(mydata['value']).mask(s).bfill()

# result
mydata['sign'] = m.diff().ne(0).cumsum()