一些错误的修复笔记

一些错误的修复笔记

第10行

df1['c_ma30'] = TA.SMA(df1,30,column='close')#原来的跟21重合了，搞错了

第13

#21日均线的斜率是否大于0，来判断是否盘整等
df1['ma21xielv_d0_flag'] = np.where(df1['ma21xielv']>=0,1,0)
df1['d21_0_count'] = ta.SUM(df1['ma21xielv_d0_flag'], timeperiod=15)

#盘平df1['d21_0_count'] = ta.SUM(df1['ma21xielv_d0_flag'], timeperiod=5)，表示的是21日均线斜率在5天内大于0的天数
df1['panping'] = np.where((df1['pan_nei_14count']>9) & (df1['d21_0_count']>2) & (abs(df1['ma21xielv'])<0.12),1,0)

#长箱体突破
df1['xttp'] = np.where((df1['panping']==1) & (df1['close']>df1['c_up_12per'])  & (df1['close']>df1['close'].shift(2)) & (df1['close']>df1['close'].shift(3)),1,0)

#盘升，后一日比前一日的收盘价高，15个里面有多个这样的，且最后一日的比较高，14日的斜率为正，比10个以上都高，盘跌类似
#中间无大跌，无大涨，不超过1次
df1['hdq_flag'] = np.where((df1['high'])>(df1['high'].shift(1)),1,0)
df1['qdh_flag'] = np.where((df1['low'])<(df1['low'].shift(1)),1,0)
df1['hdq_flag_count15'] =  ta.SUM(df1['hdq_flag'], timeperiod=15)
df1['qdh_flag_count15'] =  ta.SUM(df1['qdh_flag'], timeperiod=15)
df1['pansheng_flag'] = np.where((df1['hdq_flag_count15']>9) & (df1['d21_0_count']>9) & (df1['close']>df1['close'].shift(14)) ,1,0)
df1['pandie_flag'] = np.where((df1['qdh_flag_count15']>9) & (df1['d21_0_count']<5) & (df1['close']

df1 = df1.drop(['hdq_flag','qdh_flag','hdq_flag_count15','qdh_flag_count15','c_up_12per','c_down_12per','c_shitf1_nei_flag','c_shitf2_nei_flag','c_shitf3_nei_flag','c_shitf4_nei_flag','c_shitf5_nei_flag','c_shitf6_nei_flag','c_shitf7_nei_flag','c_shitf8_nei_flag','c_shitf9_nei_flag','c_shitf1_nei_flag','c_shitf10_nei_flag','c_shitf11_nei_flag','c_shitf12_nei_flag','c_shitf13_nei_flag','c_shitf0_nei_flag'],axis=1)

#先排序，不然shift之类的就会错
df1 = df1.sort_values(by=['ts_code','date'],ascending=True)
#重点对多头进行处理，控制多头的天数
df1['duotou'] = np.where((df1['c_ma5']-df1['c_ma21']>0)  & (df1['c_ma21']-df1['c_ma30']>0) & (df1['ma21xielv']>0),1,0)#np.where可以实现casewhen同等功能·1
df1['duotou_5'] = ta.SUM(df1['duotou'], timeperiod=5)
df1['duotou_10'] = ta.SUM(df1['duotou'], timeperiod=10)
df1['duotou_5_10'] = df1['duotou_10'] - df1['duotou_5']
df1['duotou_count'] = np.select([((df1['duotou_5_10']>0) & (df1['duotou_5']==0)),
                                  ((df1['duotou_5_10']==0) & (df1['duotou_5']==0)),
                                  ((df1['duotou_5_10']>0) & (df1['duotou_5']>0)),
                                  ((df1['duotou_5_10']==0) & (df1['duotou_5']>0))],[-1,0,1,2])
df1 = df1.drop(['duotou','duotou_5','duotou_10','duotou_5_10'],axis=1)