基于条件的Pandas递归切片及交易指标计算需求

1. 定位所有slice列值为10的行；
2. 对每个slice=10的行，确定其与前一个slice列非0值行之间的区间；
3. 计算该区间内：
   • Entry Price：open列与close列的最小值；
   • Stop Loss：low列的最小值减去10；
   • Target：Entry Price + (Entry Price - Stop Loss)*2；
4. 若前一个行的slice值也为10，则排除该行作为区间起始行。

现有DataFrame #

date      open      high       low     close  slice
0   2022-05-19 09:15:00  33461.00  33624.90  33403.20  33412.45      0
1   2022-05-19 09:20:00  33413.10  33450.65  33393.90  33429.10      0
2   2022-05-19 09:25:00  33433.20  33490.05  33421.95  33460.25      0
3   2022-05-19 09:30:00  33460.55  33509.40  33419.05  33489.80      0
4   2022-05-19 09:35:00  33492.20  33506.40  33450.30  33454.70     20
5   2022-05-19 09:40:00  33452.25  33452.95  33396.40  33436.15      0
6   2022-05-19 09:45:00  33434.30  33454.80  33401.35  33439.05      0
7   2022-05-19 09:50:00  33438.30  33482.85  33423.50  33477.30     10
8   2022-05-19 09:55:00  33480.60  33535.85  33462.40  33523.65     30
9   2022-05-19 10:00:00  33527.65  33527.65  33484.10  33521.40      0
10  2022-05-19 10:05:00  33519.35  33599.25  33505.95  33530.55      0
11  2022-05-19 10:10:00  33527.50  33544.20  33496.60  33538.65      0
12  2022-05-19 10:15:00  33540.15  33565.85  33522.75  33563.25      0
13  2022-05-19 10:20:00  33563.50  33582.45  33525.95  33539.25      0
14  2022-05-19 10:25:00  33537.25  33537.50  33511.80  33516.35     20
15  2022-05-19 10:30:00  33518.00  33561.80  33513.60  33528.55      0
16  2022-05-19 10:35:00  33527.80  33551.00  33527.55  33550.50     10
17  2022-05-19 10:40:00  33551.50  33573.60  33525.85  33537.45      0
18  2022-05-19 10:45:00  33534.80  33563.10  33510.75  33555.95      0
19  2022-05-19 10:50:00  33555.55  33573.45  33540.45  33541.00      0
20  2022-05-19 10:55:00  33545.40  33586.80  33542.75  33586.80     10

现有代码 #

def buy_sell(data):
    Time = []
    SignalBuy = [] 
    BuySL = [] 
    BuyTgt = []  

    for i in range(len(data)):
                   if data["slice"][i] == 10:
                       Time.append(data["date"][i])
                       entry = data["open"][i]
                       sl = data["low"][i] - 10
                       SignalBuy.append(entry)
                       BuySL.append(sl)
                       BuyTgt.append(entry + (entry - sl) * 2)

    return pd.Series([Time, SignalBuy, BuyTgt, BuySL])

SignalDemand = pd.DataFrame()
SignalDemand["Time"], SignalDemand["Entry Price"], SignalDemand["Target"], SignalDemand["Stop Loss"]  = buy_sell(data)

print(SignalDemand.head())

当前输出 #

Time  Entry Price    Target  Stop Loss
0  2022-05-19 09:50:00     33438.30  33487.90   33413.50
1  2022-05-19 10:35:00     33527.80  33548.30   33517.55
2  2022-05-19 10:55:00     33545.40  33570.70   33532.75

期望输出 #

Time  Entry Price    Target  Stop Loss
0  2022-05-19 09:50:00     33434.30  33530.10   33386.40
1  2022-05-19 10:35:00     33516.35  33545.45   33501.80
2  2022-05-19 10:55:00     33534.80  33602.90   33500.75

解决方案 #

修正后的代码如下，核心逻辑是找到每个slice=10行对应的有效区间，再在区间内计算指标：

import pandas as pd

def buy_sell(data):
    Time = []
    EntryPrice = []
    StopLoss = []
    Target = []
    
    # 筛选所有slice非0的行，保留索引和slice值
    non_zero_slices = data[data['slice'] != 0][['slice']].reset_index()
    # 获取所有slice=10的行的索引列表
    slice_10_indices = data[data['slice'] == 10].index.tolist()
    
    for idx in slice_10_indices:
        # 找到当前索引之前，slice非0且不等于10的最近行
        valid_prev = non_zero_slices[(non_zero_slices['index'] < idx) & (non_zero_slices['slice'] != 10)]
        if not valid_prev.empty:
            # 区间起始为前一个有效行的下一行
            start_idx = valid_prev['index'].iloc[-1] + 1
            # 提取区间内的所有行
            interval_data = data.loc[start_idx:idx]
            
            # 计算Entry Price：open和close的最小值
            entry = min(interval_data['open'].min(), interval_data['close'].min())
            # 计算Stop Loss：low的最小值减10
            sl = interval_data['low'].min() - 10
            # 计算Target
            tgt = entry + (entry - sl) * 2
            
            # 收集结果
            Time.append(data['date'].loc[idx])
            EntryPrice.append(entry)
            StopLoss.append(sl)
            Target.append(tgt)
    
    return pd.Series([Time, EntryPrice, Target, StopLoss])

# 初始化给定的DataFrame
data = pd.DataFrame({
    'date': ['2022-05-19 09:15:00', '2022-05-19 09:20:00', '2022-05-19 09:25:00', '2022-05-19 09:30:00', '2022-05-19 09:35:00', '2022-05-19 09:40:00', '2022-05-19 09:45:00', '2022-05-19 09:50:00', '2022-05-19 09:55:00', '2022-05-19 10:00:00', '2022-05-19 10:05:00', '2022-05-19 10:10:00', '2022-05-19 10:15:00', '2022-05-19 10:20:00', '2022-05-19 10:25:00', '2022-05-19 10:30:00', '2022-05-19 10:35:00', '2022-05-19 10:40:00', '2022-05-19 10:45:00', '2022-05-19 10:50:00', '2022-05-19 10:55:00'],
    'open': [33461.00, 33413.10, 33433.20, 33460.55, 33492.20, 33452.25, 33434.30, 33438.30, 33480.60, 33527.65, 33519.35, 33527.50, 33540.15, 33563.50, 33537.25, 33518.00, 33527.80, 33551.50, 33534.80, 33555.55, 33545.40],
    'high': [33624.90, 33450.65, 33490.05, 33509.40, 33506.40, 33452.95, 33454.80, 33482.85, 33535.85, 33527.65, 33599.25, 33544.20, 33565.85, 33582.45, 33537.50, 33561.80, 33551.00, 33573.60, 33563.10, 33573.45, 33586.80],
    'low': [33403.20, 33393.90, 33421.95, 33419.05, 33450.30, 33396.40, 33401.35, 33423.50, 33462.40, 33484.10, 33505.95, 33496.60, 33522.75, 33525.95, 33511.80, 33513.60, 33527.55, 33525.85, 33510.75, 33540.45, 33542.75],
    'close': [33412.45, 33429.10, 33460.25, 33489.80, 33454.70, 33436.15, 33439.05, 33477.30, 33523.65, 33521.40, 33530.55, 33538.65, 33563.25, 33539.25, 33516.35, 33528.55, 33550.50, 33537.45, 33555.95, 33541.00, 33586.80],
    'slice': [0, 0, 0, 0, 20, 0, 0, 10, 30, 0, 0, 0, 0, 0, 20, 0, 10, 0, 0, 0, 10]
})

# 生成结果DataFrame
SignalDemand = pd.DataFrame()
SignalDemand["Time"], SignalDemand["Entry Price"], SignalDemand["Target"], SignalDemand["Stop Loss"] = buy_sell(data)

print(SignalDemand)

代码说明 #

1. 先筛选出所有slice非0的行，方便快速定位每个slice=10行的前一个有效起始点；
2. 遍历每个slice=10的行索引，找到其之前最近的slice≠0且≠10的行，以此确定区间的起始位置（起始行的下一行到当前slice=10的行）；
3. 在目标区间内计算三个指标：
   • Entry Price：取区间内`