如何优雅实现 dataframe 间隔行比较不同列的值

[yellowtail]

dropmark flag high low trade_date mark markf


1680 1.0 -1.0 113.2499 111.8426 2020-06-12 09:32:00 -1 -1
1708 1.0 1.0 117.7255 117.5454 2020-06-12 10:00:00 1 1
1740 1.0 -1.0 115.7845 115.5778 2020-06-12 10:32:00 -1 -1
1871 1.0 1.0 117.3920 117.3787 2020-06-12 14:13:00 1 1
1954 1.0 -1.0 116.6383 116.2781 2020-06-15 10:06:00 -1 -1
2040 1.0 1.0 120.5869 120.3935 2020-06-15 13:02:00 1 1
2146 1.0 -1.0 118.6593 118.5726 2020-06-15 14:48:00 -1 -1
2251 1.0 1.0 120.1667 120.1534 2020-06-16 11:03:00 1 1
2383 1.0 -1.0 120.2134 120.1600 2020-06-16 14:45:00 -1 -1
2413 1.0 1.0 120.1000 120.0467 2020-06-17 09:45:00 1 1
2465 1.0 -1.0 118.0590 118.0257 2020-06-17 10:37:00 -1 -1
2474 1.0 1.0 118.6660 118.3525 2020-06-17 10:46:00 1 1
2615 1.0 -1.0 117.6321 117.5654 2020-06-17 14:37:00 -1 -1
2629 1.0 1.0 117.9790 117.8255 2020-06-17 14:51:00 1 1
2758 1.0 -1.0 117.4720 117.4187 2020-06-18 11:30:00 -1 -1
2772 1.0 1.0 117.9656 117.9256 2020-06-18 13:14:00 1 1
2869 1.0 -1.0 118.8594 118.8594 2020-06-18 14:51:00 -1 -1
2923 1.0 1.0 122.3211 122.0010 2020-06-19 10:15:00 1 1
2953 1.0 -1.0 118.7927 118.7260 2020-06-19 10:45:00 -1 -1
3063 1.0 1.0 120.4335 120.4268 2020-06-19 14:05:00 1 1
3072 1.0 -1.0 119.8466 119.7265 2020-06-19 14:14:00 -1 -1
3096 1.0 1.0 120.1267 120.0934 2020-06-19 14:38:00 1 1
3151 1.0 -1.0 114.3972 114.2438 2020-06-22 10:03:00 -1 -1
3268 1.0 1.0 118.6793 118.5793 2020-06-22 13:30:00 1 1
3287 1.0 -1.0 117.5254 117.3320 2020-06-22 13:49:00 -1 -1
3343 1.0 1.0 118.0523 117.9656 2020-06-22 14:45:00 1 1
3357 1.0 -1.0 117.4587 117.0585 2020-06-22 14:59:00 -1 -1
3376 1.0 1.0 124.9100 124.5500 2020-06-23 09:48:00 1 1
3434 1.0 -1.0 120.5100 120.5000 2020-06-23 10:46:00 -1 -1
3535 1.0 1.0 121.9800 121.9000 2020-06-23 13:57:00 1 1
3914 1.0 -1.0 115.1000 115.0500 2020-06-29 10:46:00 -1 -1
3972 1.0 1.0 116.1800 115.8500 2020-06-29 13:14:00 1 1
4077 1.0 -1.0 114.2700 114.2600 2020-06-29 14:59:00 -1 -1
4180 1.0 1.0 121.8400 121.6100 2020-06-30 11:12:00 1 1
4209 1.0 -1.0 122.2600 121.8700 2020-06-30 13:11:00 -1 -1
4236 1.0 1.0 123.7000 123.4100 2020-06-30 13:38:00 1 1
4245 1.0 -1.0 122.5500 122.3300 2020-06-30 13:47:00 -1 -1

要求找出是 markf 列值为-1 的值大于下一个行的 high 列值，用 for 循环感觉效率太低了...

[czwstc]

最后一句话我读了大概十几遍没看懂。
你要么描述一下？

[cassidyhere]

df['diff'] = df.high.shift() - df.high
df[df.diff > 0][df.markf == -1]

[14cheese]

DataFrame.shift 方法了解一下

[liangch]

shift(1) shift(-1)随你玩。

[yellowtail]

@czwstc 少打了一个 low 值...要求找出所有 markf 列值为-1 的，且大于下一个行的 high 列值，的 low 值