Pandas 通过跳过一行并在不同值上重置来计数的计数器
Pandas counter that counts by skipping a row and reset on different values
您好，我正在尝试创建一个计数器来计算我的趋势列，方法是跳过一行并在字符串值不同时自行重置。例如，在第 9 行，它会计数 2，因为前一个跳过的行被计数为 1。但它会重置回 1，因为第 11 行的值与第 9 行不同。
无论如何我可以做到这一点吗？
DateTimeStarted 50% Quantile 50Q shift 2H Trend Count
0 2020-12-18 15:00:00 554.0 NaN Flat 1
1 2020-12-18 16:00:00 593.0 NaN Flat 1
2 2020-12-18 17:00:00 534.0 554.0 Down 1
3 2020-12-18 18:00:00 562.0 593.0 Down 1
4 2020-12-18 19:00:00 552.0 534.0 Up 1
5 2020-12-18 20:00:00 592.0 562.0 Up 1
6 2020-12-19 08:00:00 511.0 552.0 Down 1
7 2020-12-19 09:00:00 584.0 592.0 Down 1
8 2020-12-19 10:00:00 576.0 511.0 Up 1
9 2020-12-19 11:00:00 545.5 584.0 Down 2
10 2020-12-19 12:00:00 609.5 576.0 Up 2
11 2020-12-19 13:00:00 548.0 545.5 Up 1
12 2020-12-19 14:00:00 565.0 609.5 Down 1
13 2020-12-19 15:00:00 575.0 548.0 Up 2
14 2020-12-19 16:00:00 570.0 565.0 Up 1
15 2020-12-19 17:00:00 557.0 575.0 Down 1
16 2020-12-19 18:00:00 578.0 570.0 Up 2
17 2020-12-19 19:00:00 578.5 557.0 Up 1
18 2020-12-21 08:00:00 543.0 578.0 Down 1
19 2020-12-21 09:00:00 558.0 578.5 Down 1
20 2020-12-21 10:00:00 570.0 543.0 Up 1
您可以 shift() Trend 列乘以 2 并检查它是否等于 Trend:
df['Counter'] = df.Trend.shift(2).eq(df.Trend).astype(int).add(1)
这里我给它取名为Counter，方便比较：
DateTimeStarted 50%Quantile 50Qshift2H Trend Count Counter
0 2020-12-18 15:00:00 554.0 NaN Flat 1 1
1 2020-12-18 16:00:00 593.0 NaN Flat 1 1
2 2020-12-18 17:00:00 534.0 554.0 Down 1 1
3 2020-12-18 18:00:00 562.0 593.0 Down 1 1
4 2020-12-18 19:00:00 552.0 534.0 Up 1 1
5 2020-12-18 20:00:00 592.0 562.0 Up 1 1
6 2020-12-19 08:00:00 511.0 552.0 Down 1 1
7 2020-12-19 09:00:00 584.0 592.0 Down 1 1
8 2020-12-19 10:00:00 576.0 511.0 Up 1 1
9 2020-12-19 11:00:00 545.5 584.0 Down 2 2
10 2020-12-19 12:00:00 609.5 576.0 Up 2 2
11 2020-12-19 13:00:00 548.0 545.5 Up 1 1
12 2020-12-19 14:00:00 565.0 609.5 Down 1 1
13 2020-12-19 15:00:00 575.0 548.0 Up 2 2
14 2020-12-19 16:00:00 570.0 565.0 Up 1 1
15 2020-12-19 17:00:00 557.0 575.0 Down 1 1
16 2020-12-19 18:00:00 578.0 570.0 Up 2 2
17 2020-12-19 19:00:00 578.5 557.0 Up 1 1
18 2020-12-21 08:00:00 543.0 578.0 Down 1 1
19 2020-12-21 09:00:00 558.0 578.5 Down 1 1
20 2020-12-21 10:00:00 570.0 543.0 Up 1 1
您好，我正在尝试创建一个计数器来计算我的趋势列，方法是跳过一行并在字符串值不同时自行重置。例如，在第 9 行，它会计数 2，因为前一个跳过的行被计数为 1。但它会重置回 1，因为第 11 行的值与第 9 行不同。 无论如何我可以做到这一点吗？
DateTimeStarted 50% Quantile 50Q shift 2H Trend Count
0 2020-12-18 15:00:00 554.0 NaN Flat 1
1 2020-12-18 16:00:00 593.0 NaN Flat 1
2 2020-12-18 17:00:00 534.0 554.0 Down 1
3 2020-12-18 18:00:00 562.0 593.0 Down 1
4 2020-12-18 19:00:00 552.0 534.0 Up 1
5 2020-12-18 20:00:00 592.0 562.0 Up 1
6 2020-12-19 08:00:00 511.0 552.0 Down 1
7 2020-12-19 09:00:00 584.0 592.0 Down 1
8 2020-12-19 10:00:00 576.0 511.0 Up 1
9 2020-12-19 11:00:00 545.5 584.0 Down 2
10 2020-12-19 12:00:00 609.5 576.0 Up 2
11 2020-12-19 13:00:00 548.0 545.5 Up 1
12 2020-12-19 14:00:00 565.0 609.5 Down 1
13 2020-12-19 15:00:00 575.0 548.0 Up 2
14 2020-12-19 16:00:00 570.0 565.0 Up 1
15 2020-12-19 17:00:00 557.0 575.0 Down 1
16 2020-12-19 18:00:00 578.0 570.0 Up 2
17 2020-12-19 19:00:00 578.5 557.0 Up 1
18 2020-12-21 08:00:00 543.0 578.0 Down 1
19 2020-12-21 09:00:00 558.0 578.5 Down 1
20 2020-12-21 10:00:00 570.0 543.0 Up 1
您可以 shift() Trend 列乘以 2 并检查它是否等于 Trend:
df['Counter'] = df.Trend.shift(2).eq(df.Trend).astype(int).add(1)
这里我给它取名为Counter，方便比较：
DateTimeStarted 50%Quantile 50Qshift2H Trend Count Counter
0 2020-12-18 15:00:00 554.0 NaN Flat 1 1
1 2020-12-18 16:00:00 593.0 NaN Flat 1 1
2 2020-12-18 17:00:00 534.0 554.0 Down 1 1
3 2020-12-18 18:00:00 562.0 593.0 Down 1 1
4 2020-12-18 19:00:00 552.0 534.0 Up 1 1
5 2020-12-18 20:00:00 592.0 562.0 Up 1 1
6 2020-12-19 08:00:00 511.0 552.0 Down 1 1
7 2020-12-19 09:00:00 584.0 592.0 Down 1 1
8 2020-12-19 10:00:00 576.0 511.0 Up 1 1
9 2020-12-19 11:00:00 545.5 584.0 Down 2 2
10 2020-12-19 12:00:00 609.5 576.0 Up 2 2
11 2020-12-19 13:00:00 548.0 545.5 Up 1 1
12 2020-12-19 14:00:00 565.0 609.5 Down 1 1
13 2020-12-19 15:00:00 575.0 548.0 Up 2 2
14 2020-12-19 16:00:00 570.0 565.0 Up 1 1
15 2020-12-19 17:00:00 557.0 575.0 Down 1 1
16 2020-12-19 18:00:00 578.0 570.0 Up 2 2
17 2020-12-19 19:00:00 578.5 557.0 Up 1 1
18 2020-12-21 08:00:00 543.0 578.0 Down 1 1
19 2020-12-21 09:00:00 558.0 578.5 Down 1 1
20 2020-12-21 10:00:00 570.0 543.0 Up 1 1