学习笔记 select的用法_df1['stock qty'] = np.where(df1['special procureme-优快云博客

本文链接：https://blog.youkuaiyun.com/sphinxrascal168/article/details/122211369

本文探讨了如何通过分段对股票当日涨跌进行标记，涉及百分比变化阈值的设置，并结合技术指标如连续十字星、扁线判断方法，揭示了交易信号的生成过程。同时，关注了K_bianxian_flag的计算及其在形态识别中的应用。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

#####对当日涨跌进行分段
df1['pct_chg_cate'] = np.select([((df1['pct_chg']>=4.9999)),
((df1['pct_chg'] <4.9999) & (df1['pct_chg']>=1.5001)),
((df1['pct_chg'] <(1.5001)) & (df1['pct_chg']>=(-1.5001))),
((df1['pct_chg'] <(-1.5001)) & (df1['pct_chg']>=(-4.9999))),((df1['pct_chg'] <(-4.9999)))],[5,4,3,2,1])

#连续十字星
df1['shizi_panzheng_1'] = ta.SUM(df1['K_bianxian_flag'], timeperiod=15)
df1['shizi_panzheng'] = np.where((df1['shizi_panzheng_1']>9) & (abs(df1['ma89xielv'])<1.2) & (df1['ma21xielv']>0)),1,0)

#是否扁线（实体部分绝对长度低，且当天涨幅小，要修改，上次是根据历史长度来确定扁线）
df1['K_bianxian_flag'] = np.where(df1['shiti_len']<=1.5,1,0)

#是否扁线（实体部分绝对长度低，且当天涨幅小）
df1['shiti_len_count10'] = ta.SUM(df1['shiti_len'], timeperiod=10)
df1['shiti_len_count30'] = ta.SUM(df1['shiti_len'], timeperiod=30)
df1['shiti_len_pingjun'] = (df1['shiti_len_count30'] - df1['shiti_len_count10'])/60
df1['K_bianxian_flag'] = np.where(df1['shiti_len']<=1.5 & df1['shiti_len']<df1['shiti_len_pingjun'],1,0)
df1 = df1.drop(['shiti_len_count10','shiti_len_count30','shiti_len_pingjun'],axis=1)