根据时间查询放款集，申请集

data/samples/xyqbdata.py

@@ -6,15 +6,20 @@ from data.samples import mysqlquery
 '''
 
 
-def query_user_loan_type(order_nos):
+def query_risk_info(order_nos):
     '''
+    直接使用了线上的用户类型，而非分析库的计算得到的用户类型
     :param order_nos:
-    :return:order_no,user_loan_type-- 策略使用的用户类型
+    :return:order_no,user_loan_type-- 策略使用的用户类型,审核拒绝原因，是否黑名单拒绝或其他拒绝
     '''
     res = []
     sql = '''
     select biz_no as order_no,
-    ifnull(JSON_EXTRACT(audit_context_data, '$.user_loan_type_v3'),JSON_EXTRACT(audit_context_data, '$.user_loan_type_v4')) as user_loan_type
+    ifnull(JSON_EXTRACT(audit_context_data, '$.user_loan_type_v3'),
+    JSON_EXTRACT(audit_context_data, '$.user_loan_type_v4')) as user_loan_type,
+    JSON_EXTRACT(audit_context_data, '$.filter_info_in_black') qg_black,
+    JSON_EXTRACT(audit_context_data, '$."third_data_source#xy_neg_code"') as xy_black,
+    audit_result
     from biz_audit_log_data 
     where biz_no in %s 
     '''
@@ -23,7 +28,11 @@ def query_user_loan_type(order_nos):
         res.append(pd.read_sql(sql % str(tuple(order_nos[i:i + 1000])), engine_risk))
     df= pd.concat(res)
     df.drop_duplicates(['order_no'],inplace=True)
-    return df
+    df['refuse']='其他'
+    df.loc[df.audit_result==1,'refuse']='审核通过'
+    df.loc[(df.audit_result!=1) & (df.qg_black> 0),'refuse']='命中QG黑名单'
+    df.loc[(df.audit_result != 1) & (df.xy_black=='0'), 'refuse'] = '命中新颜黑名单'
+    return df[['order_no','uesr_loan_type','refuse']]
 
 def query_byloanid(loan_ids):
     '''
@@ -54,8 +63,8 @@ def query_byloanid(loan_ids):
 def query_by_orderno(order_nos):
     '''
         数据源为分析库
-        :param loan_ids:list 放款集
-        :return:order_no,user_id,loan_id,用户类型,渠道，申请时间，拒绝原因，首逾天数，最大逾期天数，放款期数
+        :param order_nos:list 放款集
+        :return:order_no,user_id,loan_id,用户类型,渠道，申请时间，拒绝原因，首逾天数，最大逾期天数，放款期数 reason['已放款'，'审核通过','审核未通过','黑名单']
         '''
     sql = '''
         select t1.loan_id,t1.user_id,t2.order_no,t2.contract_term,if(t2.term_no==1,t2.passdue_day,null) as passdue_day,
@@ -75,4 +84,44 @@ def query_by_orderno(order_nos):
     # == 剔除重复数据
     df.sort_values(['loan_id'], ascending=True, inplace=True)
     df.drop_duplicates(['loan_id'], keep='last', inplace=True)
-    return df
+    return df
\ No newline at end of file
+
+
+def query_bydate(start_date,end_date,is_loan=True):
+    '''
+    获取指定日期的数据,如果 is_loan=True ,则为放款集 否则为申请集
+    :param start_date:
+    :param end_date:
+    :return:dataframe
+    '''
+    if is_loan:
+        sql='''
+        select t1.loan_id,t1.user_id,t2.order_no,t2.contract_term,if(t2.term_no==1,t2.passdue_day,null) as passdue_day,
+        t1.applied_at,t1.applied_type,t1.applied_channel,if(t2.loan_id is not null,'已放款',if(t1.approval==1,'审核通过','审核未通过')) as reason,
+        max(t2.passdue_day) as max_passdue_day
+        from loan_application t1 
+        join loan_repay t2 on t1.loan_id=t2.loan_id and t2.repayment_status!=4
+        where  t1.applied=1
+        and t1.applied_at >= '%s' and t1.applied_at < '%s'
+        group by 1,2,3,4,5,6,7,8
+        ''' % (start_date,end_date)
+    else:
+        sql='''
+        select t1.loan_id,t1.user_id,t2.order_no,t2.contract_term,if(t2.term_no==1,t2.passdue_day,null) as passdue_day,
+        t1.applied_at,t1.applied_type,t1.applied_channel,if(t2.loan_id is not null,'已放款',if(t1.approval==1,'审核通过','审核未通过')) as reason,
+        max(t2.passdue_day) as max_passdue_day
+        from loan_application t1 
+        left join loan_repay t2 on t1.loan_id=t2.loan_id and t2.repayment_status!=4
+        where  t1.applied=1
+        and t1.applied_at >= '%s' and t1.applied_at < '%s'
+        group by 1,2,3,4,5,6,7,8
+        ''' % (start_date,end_date)
+
+    df = mysqlquery.query(sql,engine_analysis_new)
+    df.sort_values(['loan_id'],ascending=True,inplace=True)
+    df.drop_duplicates(['loan_id'],keep='last',inplace=True)
+    tmp=query_risk_info(df.order_no.tolist())
+    return pd.merge(df,tmp,on='order_no',how='left')
+
+
+