rt

审计第一阶段/现金贷/收入确认表-不同库版本.py

@@ -25,6 +25,8 @@ pd.options.mode.chained_assignment = None
 pd.set_option('display.float_format', lambda x: '%.5f' % x)
 
 res_sum = []
+
+
 def proc_save(watch_date):
     loan = df_loan_ratio.loc[df_loan_ratio[u'放款日期'] < watch_date]
     loan.loc[~loan.is_active.isin([1, -2]), u'本金债权'] = 0
@@ -84,12 +86,16 @@ def proc_save(watch_date):
                                         0)
     print(
         loan_result[
-            [u'本金债权', u'计提风险金', u'qg_one_time_service_fee', u'实还服务费（减免后）', u'趸交服务费', u'实还服务费（减免后、含一次性服务费、趸交服务费）', u'asset余额',
+            [u'本金债权', u'计提风险金', u'qg_one_time_service_fee', u'实还服务费（减免后）', u'趸交服务费', u'实还服务费（减免后、含一次性服务费、趸交服务费）',
+             u'asset余额',
              u'收入确认金额']].sum())
     tmp_sum = loan_result[
         [u'本金债权', u'计提风险金', u'qg_one_time_service_fee', u'实还服务费（减免后）', u'趸交服务费', u'实还服务费（减免后、含一次性服务费、趸交服务费）',
          u'asset余额',
          u'收入确认金额']].sum().to_frame().T
+    # tmp_sum = loan_result.groupby(u'风险等级')[
+    #     [u'本金债权', u'计提风险金', u'qg_one_time_service_fee', u'实还服务费（减免后）', u'趸交服务费', u'实还服务费（减免后、含一次性服务费、趸交服务费）',
+    #      u'asset余额', u'收入确认金额']].agg('sum').reset_index()
     tmp_sum[u'截止时间'] = watch_date
     res_sum.append(tmp_sum)
 
@@ -107,7 +113,7 @@ def proc_save(watch_date):
 if __name__ == '__main__':
     step = 10000
     max_limit = 900000
-    file_path = u'E:/审计出表/现金贷/0926'
+    file_path = u'E:/审计出表/现金贷/1021'
     file_name = u'现金贷收入确认表'
     if not os.path.exists(file_path):
         os.makedirs(file_path)
@@ -129,9 +135,12 @@ if __name__ == '__main__':
     df_ms.loan_id = df_ms.loan_id.astype(str)
 
     db_names = ['all_back_0630', 'all_back_0731', 'all_back_0831', 'new_transaction']
-    date_report_list = [[item.date() for item in pd.date_range(start='2017-02', end='2017-07', freq='MS')],
-                        [datetime.date(2017, 8, 1)],
-                        [datetime.date(2017, 9, 1)], [datetime.date(2017, 10, 1)]]
+    date_report_list = [
+        [datetime.date(2017, 2, 1), datetime.date(2017, 3, 1), datetime.date(2017, 4, 1), datetime.date(2017, 5, 1),
+         datetime.date(2017, 6, 1), datetime.date(2017, 7, 1)],
+        [datetime.date(2017, 8, 1)],
+        [datetime.date(2017, 9, 1)],
+        [datetime.date(2017, 10, 1)]]
     # db_names = ['new_transaction',]
     # date_report_list = [[datetime.date(2017, 10, 1)],]
     for i in xrange(len(db_names)):
@@ -141,7 +150,8 @@ if __name__ == '__main__':
         # #######################
         # 查文件
         df_loan = pd.read_csv(os.path.join(file_path, 'all_loan_%s.csv' % db_name), encoding='utf8', low_memory=False)
-        df_ref = pd.read_csv(os.path.join(file_path, 'all_ref_%s.csv' % db_name), encoding='utf8', low_memory=False)
+        df_ref = pd.read_csv(os.path.join(file_path, 'all_ref_%s.txt' % db_name), sep='\t', low_memory=False)
+        df_ref.columns = [u'ref_id', u'实还总额', u'实还本金', u'实还利息', u'实还服务费（减免后）', u'实还罚息', u'其他收益', u'repaid_at']
 
         df_loan[u'放款日期'] = pd.to_datetime(df_loan[u'放款日期']).dt.date
         df_loan.loc[df_loan[u'风险等级'] == 'D', u'风险等级'] = 'C'

审计第一阶段/现金贷/收入确认表-读数据.py

@@ -25,9 +25,9 @@ step = 10000
 max_limit = 900000
 pd.options.mode.chained_assignment = None
 pd.set_option('display.float_format', lambda x: '%.5f' % x)
-db_names = ['all_back_0630', 'all_back_0731', 'all_back_0831']
-end_dates = [datetime.date(2017, 7, 1), datetime.date(2017, 8, 1), datetime.date(2017, 9, 1)]
-file_path = u'E:/审计出表/现金贷/0926'
+db_names = ['audit_temp']
+end_dates = [datetime.date(2017, 7, 1)]
+file_path = u'E:/审计出表/现金贷/1021'
 
 if not os.path.exists(file_path):
     os.makedirs(file_path)
@@ -66,7 +66,7 @@ SELECT
   xrprrr.punish - xrprrr.mitigate_collection_relief 实还罚息,
   xrprrr.remain_income                              其他收益,
   DATE(xrprrrrd.repaid_at)                          `repaid_at`
-FROM %s.xjd_repay_plan_repay_record_ref xrprrr
+FROM %s.xjd_repay_plan_repay_record_ref_0630 xrprrr
   JOIN test_mzh.xjd_repay_plan_repay_record_ref_repay_date_flow xrprrrrd ON xrprrr.id = xrprrrrd.xjd_ref_id
   JOIN %s.user_repayment_plan urp ON urp.id = xrprrr.plan_id
 WHERE xrprrrrd.repaid_at < '%s';
@@ -116,13 +116,13 @@ for i in xrange(0, len(db_names)):
     end_date = end_dates[i]
     # ##############################################
     # 查数据库，并保存文件，时间较长，服务器上在15分钟左右，本机25分钟左右
-    df_gen = pd.read_sql(sql=sql_loan_plan % (db_name, end_date,), con=mysql, chunksize=step)
-    res = []
-    for tmp in df_gen:
-        res.append(tmp)
-    df_loan = pd.concat(res)
-    df_loan['qg_one_time_service_fee'].fillna(0, inplace=True)
-    df_loan.to_csv(os.path.join(file_path, 'all_loan_%s.csv' % db_name), encoding='utf8', index=None)
+    # df_gen = pd.read_sql(sql=sql_loan_plan % (db_name, end_date,), con=mysql, chunksize=step)
+    # res = []
+    # for tmp in df_gen:
+    #     res.append(tmp)
+    # df_loan = pd.concat(res)
+    # df_loan['qg_one_time_service_fee'].fillna(0, inplace=True)
+    # df_loan.to_csv(os.path.join(file_path, 'all_loan_%s.csv' % db_name), encoding='utf8', index=None)
 
     df_gen = pd.read_sql(sql=sql_ref % (db_name, db_name, end_date,), con=mysql, chunksize=step)
     res = []
@@ -131,9 +131,9 @@ for i in xrange(0, len(db_names)):
     df_ref = pd.concat(res)
     df_ref.to_csv(os.path.join(file_path, 'all_ref_%s.csv' % db_name), encoding='utf8', index=None)
 
-df_ms = pd.read_sql(sql=sql_ms, con=mysql)
-df_ms.to_csv(os.path.join(file_path, 'ms.csv'), encoding='utf8', index=None)
-df_yghs = pd.read_sql(sql=sql_yghs, con=mysql)
-df_yghs.to_csv(os.path.join(file_path, 'yghs.csv'), encoding='utf8', index=None)
-df_dunjiao = pd.read_sql(sql=sql_dunjiao, con=mysql)
-df_dunjiao.to_csv(os.path.join(file_path, 'dunjiao.csv'), encoding='utf8', index=None)
+# df_ms = pd.read_sql(sql=sql_ms, con=mysql)
+# df_ms.to_csv(os.path.join(file_path, 'ms.csv'), encoding='utf8', index=None)
+# df_yghs = pd.read_sql(sql=sql_yghs, con=mysql)
+# df_yghs.to_csv(os.path.join(file_path, 'yghs.csv'), encoding='utf8', index=None)
+# df_dunjiao = pd.read_sql(sql=sql_dunjiao, con=mysql)
+# df_dunjiao.to_csv(os.path.join(file_path, 'dunjiao.csv'), encoding='utf8', index=None)

审计第一阶段/白条/去哪儿/去哪儿收入确认表-模型.py

@@ -100,7 +100,7 @@ def cal_debt_age(x):
 
 df_res['账龄'] = df_res[['放款时间', '应还款日']].apply(lambda x: cal_debt_age(x), axis=1)
 
-asset_remain = df_res.loc[df_res.assets_remain > 0]
+asset_remain = df_res.loc[df_res.assets_remain != 0]
 asset_remain[['产品编号', '风险等级', '放款时间', '应还服务费', 'assets_remain']].to_excel(
     os.path.join(file_path, '去哪儿asset_remain.xlsx'),
     index=None)
@@ -111,12 +111,12 @@ gp1.to_excel(os.path.join(file_path, '去哪儿汇总（期数-应还年月-风
 gp2 = df_res.groupby(['账龄', '应还年月', '风险等级'])['assets'].agg('sum').reset_index()
 gp2.to_excel(os.path.join(file_path, '去哪儿汇总（账龄-应还年月-风险等级）.xlsx'), index=None)
 
-# df_res.to_csv(os.path.join(file_path, '去哪儿assets减值-details.csv'), index=None, encoding='gbk')
+df_res.to_csv(os.path.join(file_path, '去哪儿assets减值-details.csv'), index=None, encoding='gbk')
 #
 # max_limit = 900000
 # if len(df_res) > max_limit:
 #     for i in xrange(0, len(df_res), max_limit):
 #         df_res[i:i + max_limit].to_csv(os.path.join(file_path, '去哪儿assets减值-details_%d.csv' % (i / max_limit + 1)), index=None,
 #                                        encoding='utf8')
-# df_res.loc[df_res['产品编号'].duplicated(), '放款金额'] = 0
-# print('{}'.format(df_res['放款金额'].sum()))
+df_res.loc[df_res['产品编号'].duplicated(), '放款金额'] = 0
+print('{}'.format(df_res['放款金额'].sum()))

审计第一阶段/白条/去哪儿/去哪儿收入确认表.py

@@ -90,7 +90,7 @@ df_loan['计提风险金'] = 0
 
 view_dates = [datetime.date(2017, 9, 1), datetime.date(2017, 10, 1)]
 file_path = 'E:/审计出表/白条/收入确认表'
-
+res_sum = []
 for v_date in view_dates:
     last_day = v_date + datetime.timedelta(days=-1)
     df_loan_v = df_loan.loc[df_loan['放款时间'] < v_date]
@@ -141,7 +141,13 @@ for v_date in view_dates:
     df_loan_v.drop(['减值', '退款服务费'], axis=1, inplace=True)
     df_loan_v['放款时间'] = df_loan_v['放款时间'].dt.date
     print(df_loan_v[['放款金额', '计提风险金', '实还服务费', 'assets', '收入']].sum())
-    print(df_loan_v.groupby(['风险等级']).agg('sum'))
+    # tmp_sum = df_loan_v.groupby(u'风险等级')[
+    #     [u'放款金额', u'计提风险金', u'实还服务费', u'assets', u'收入']].agg('sum').reset_index()
+    tmp_sum = df_loan_v[[u'放款金额', u'计提风险金', u'实还服务费', u'assets', u'收入']].agg('sum').to_frame().T
+    tmp_sum[u'截止时间'] = v_date
+    res_sum.append(tmp_sum)
     df_loan_v.to_csv(os.path.join(file_path, '白条去哪儿收入确认表_{}_{:0>2}.csv'.format(v_date.year, v_date.month - 1)),
                      index=None,
                      encoding='gbk')
+df_sum = pd.concat(res_sum)
+df_sum.to_excel(os.path.join(file_path, '白条去哪儿汇总.xlsx'), index=None)
\ No newline at end of file

审计第一阶段/白条/非去哪儿/非去哪儿收入确认表-模型.py

@@ -149,9 +149,9 @@ def cal_debt_age(x):
 
 df_res['账龄'] = df_res[['放款时间', '应还日']].apply(lambda x: cal_debt_age(x), axis=1)
 
-# asset_remain = df_res.loc[df_res.assets_remain > 0]
-# asset_remain[['产品编号', '风险等级', '放款时间', '应还服务费', 'assets_remain']].to_excel(os.path.join(file_path, '非去哪儿asset_remain.xlsx'),
-#                                                                           index=None)
+asset_remain = df_res.loc[df_res.assets_remain != 0]
+asset_remain[['产品编号', '风险等级', '放款时间', '应还服务费', 'assets_remain']].to_excel(os.path.join(file_path, '非去哪儿asset_remain.xlsx'),
+                                                                          index=None)
 
 df_res['应还年月'] = df_res['应还日'].apply(lambda x: x.strftime('%Y-%m'))
 gp1 = df_res.groupby(['当前期数', '应还年月', '风险等级'])['assets'].agg('sum').reset_index()
@@ -160,7 +160,7 @@ gp1.to_excel(os.path.join(file_path, '非去哪儿汇总（期数-应还年月-
 gp2 = df_res.groupby(['账龄', '应还年月', '风险等级'])['assets'].agg('sum').reset_index()
 gp2.to_excel(os.path.join(file_path, '非去哪儿汇总（账龄-应还年月-风险等级）.xlsx'), index=None)
 #
-# df_res.to_csv(os.path.join(file_path, '非去哪儿assets减值-details.csv'), index=None, encoding='gbk')
+df_res.to_csv(os.path.join(file_path, '非去哪儿assets减值-details.csv'), index=None, encoding='gbk')
 #
 # max_limit = 900000
 # for i in xrange(0, len(df_res), max_limit):

审计第一阶段/白条/非去哪儿/非去哪儿收入确认表.py

@@ -102,6 +102,7 @@ print(df_repay.groupby(df_repay['还款时间'].dt.month)['实还服务费'].agg
 
 view_dates = [datetime.date(2017, 9, 1), datetime.date(2017, 10, 1)]
 file_path = 'E:/审计出表/白条/收入确认表'
+res_sum = []
 for v_date in view_dates:
     df_loan_v = df_loan.loc[df_loan['放款时间'] < v_date]
     df_repay_v = df_repay.loc[df_repay['还款时间'] < v_date]
@@ -132,8 +133,17 @@ for v_date in view_dates:
     df_loan_v.drop(['ref_id', '减值'], axis=1, inplace=True)
     df_loan_v['放款时间'] = df_loan_v['放款时间'].dt.date
     print(df_loan_v[['放款金额', '计提风险金', '实还服务费', 'assets', '收入']].sum())
-    # df_loan_v.to_csv(os.path.join(file_path, '白条收入确认表_{}_{:0>2}.csv'.format(v_date.year, v_date.month - 1)), index=None, encoding='gbk')
 
+    # tmp_sum = df_loan_v.groupby(u'风险等级')[
+    #     [u'放款金额', u'计提风险金', u'实还服务费', u'assets', u'收入']].agg('sum').reset_index()
+    tmp_sum = df_loan_v[[u'放款金额', u'计提风险金', u'实还服务费', u'assets', u'收入']].agg('sum').to_frame().T
+    tmp_sum[u'截止时间'] = v_date
+    res_sum.append(tmp_sum)
+    df_loan_v.to_csv(os.path.join(file_path, '白条收入确认表_{}_{:0>2}.csv'.format(v_date.year, v_date.month - 1)), index=None,
+                     encoding='gbk')
+
+df_sum = pd.concat(res_sum)
+df_sum.to_excel(os.path.join(file_path, '白条非去哪儿汇总.xlsx'), index=None)
 
 # 98108, 98108
 # ()
@@ -144,4 +154,4 @@ for v_date in view_dates:
 # 0     8   8561.64000
 # 1     9 559489.81000
 # 有还款，没有放款的len: 0，ref_id: ()
-# 有还款，没有放款的len: 0，ref_id: ()
\ No newline at end of file
+# 有还款，没有放款的len: 0，ref_id: ()