加入业务数据

data/samples/mysqlquery.py

+import pandas as pd
+
+def query(sql,engine_sql):
+    '''
+    查询大量数据
+    :param sql:
+    :param engine_sql:查询器
+    :return:dataframe
+    '''
+    res=[]
+    tmp=pd.read_sql(sql,engine_sql,chunksize=10000)
+    for tt in tmp:
+        res.append(tt)
+    return pd.concat(res)
\ No newline at end of file

data/samples/xyqbdata.py

+import pandas as pd
+from data.datasource.mysqldb import *
+from data.samples import mysqlquery
+'''
+目的：提供业务数据，包括 order_no,loan_id,用户类型【直接使用risk_info】,渠道，申请时间，拒绝原因，首逾天数，最大逾期天数，放款期数
+'''
+
+
+def query_user_loan_type(order_nos):
+    '''
+    :param order_nos:
+    :return:order_no,user_loan_type-- 策略使用的用户类型
+    '''
+    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
+    from biz_audit_log_data 
+    where biz_no in %s 
+    '''
+    for i in range(0, len(order_nos), 1000):
+        print('----exe sql %d---- ' % i)
+        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
+
+def query_byloanid(loan_ids):
+    '''
+    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
+    '''
+    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.loan_id in %s
+    group by 1,2,3,4,5,6,7,8
+    '''
+    res=[]
+    for i in range(0, len(loan_ids), 1000):
+        print('----exe sql %d---- ' % i)
+        res.append(pd.read_sql(sql % str(tuple(loan_ids[i:i + 1000])), engine_analysis_new))
+    df = pd.concat(res)
+    df.order_no = df.order_no.apply(lambda x:x.decode('utf8'))
+    #== 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
+
+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
+        '''
+    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.order_no in %s
+        group by 1,2,3,4,5,6,7,8
+        '''
+    res = []
+    for i in range(0, len(order_nos), 1000):
+        print('----exe sql %d---- ' % i)
+        res.append(pd.read_sql(sql % str(tuple(order_nos[i:i + 1000])), engine_analysis_new))
+    df = pd.concat(res)
+    df.order_no = df.order_no.apply(lambda x: x.decode('utf8'))
+    # == 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
\ No newline at end of file