因子分享-日内主动买入强度占比因子
由bqtnziby创建，最终由bqtnziby 被浏览 5 用户
1.主动买入定义及因子逻辑
主动买入金额：当分钟内收盘价 > 开盘价时，认为该分钟的成交由主动买入主导，此时主动买入金额等于该分钟的成交额；否则为 0。
相应的，被动买入金额即为 当分钟内收盘价 < 开盘价
当分钟内收盘价 > 开盘价时，认为买方更主动（愿意以更高价格买入），这类成交被归类为 “主动买入”。主动买入占比越高，说明当日买方主动进攻的力量越强。
主动买入占比：
（1）计算当日主动买入总金额
（2）计算当日总买入金额
（3）主动买入占比=主动买入总金额/总买入金额
2.次日相对涨跌幅
次日相对涨跌幅 =（次日收盘价 - 当日收盘价）/ 当日收盘价
3.最终因子
最终因子=主动买入占比*次日相对涨跌幅
因子通过乘以 “次日相对涨跌幅”，将主动买入力量与短期收益关联。逻辑是：若主动买入力量强（占比高）且次日确实上涨（涨跌幅为正），则因子值高，代表 “主动买入力量有效转化为后续收益”；反之，若主动买入力量强但次日下跌，因子值会较低甚至为负。
4.核心部分代码
sql = f"""
WITH minute_active_passive AS (
SELECT
trading_day,
instrument,
date,
amount,
close,
open,
CASE WHEN close > open THEN amount ELSE 0 END AS active_buy_amount,
CASE WHEN close < open THEN amount ELSE 0 END AS passive_buy_amount
FROM {datasource}
WHERE close > 0 AND amount > 0
),
daily_agg AS (
SELECT
trading_day,
instrument,
SUM(active_buy_amount) AS sum_active,
SUM(active_buy_amount + passive_buy_amount) AS sum_total,
MAX(date) AS last_dt
FROM minute_active_passive
GROUP BY trading_day, instrument
HAVING SUM(active_buy_amount + passive_buy_amount) > 0
),
daily_active_ratio AS (
SELECT
a.trading_day,
a.instrument,
CAST(a.sum_active AS DOUBLE) / NULLIF(a.sum_total, 0) AS active_buy_ratio,
mc.close AS day_close
FROM daily_agg a
LEFT JOIN minute_active_passive mc
ON mc.trading_day = a.trading_day AND mc.instrument = a.instrument AND mc.date = a.last_dt
),
factor_result AS (
SELECT
STRPTIME(CAST(trading_day AS VARCHAR), '%Y%m%d') AS date,
instrument,
active_buy_ratio * ((LEAD(day_close) OVER (PARTITION BY instrument ORDER BY trading_day) - day_close) / NULLIF(day_close, 0)) AS factor
FROM daily_active_ratio
)
SELECT date, instrument, factor FROM factor_result WHERE factor IS NOT NULL
"""
lookback_days = 3
query_start_date = pd.to_datetime(start_date) - pd.Timedelta(days=lookback_days)
df = dai.query(sql, filters={'date': [query_start_date, end_date]}, compression=True).df()
df = df[df['date'].between(start_date, end_date)]
return df
\