update

new_user_analysis/user_behavior_path.py

@@ -85,7 +85,6 @@ spark.sql("CREATE TEMPORARY FUNCTION json_map AS 'brickhouse.udf.json.JsonMapUDF
 spark.sql("CREATE TEMPORARY FUNCTION is_json AS 'com.gmei.hive.common.udf.UDFJsonFormatCheck'")
 spark.sql("CREATE TEMPORARY FUNCTION arrayMerge AS 'com.gmei.hive.common.udf.UDFArryMerge'")
 
-
 task_list = []
 task_days = 30
 
@@ -205,14 +204,14 @@ name_dict = {
     "diary_card_click_vote": "日记本卡片点赞或取消赞",
     "diary_detail_bottom_click_same_service": "日记贴详情页点击同类美购按钮",
     "diary_detail_bottom_click_samediary_icon": "日记本详情页_底栏_点击_同类日记按钮",
-      "diary_item_click_avatar": "日记本卡片 - 点击作者头像",
-      "diary_item_click_banner": "首页精选tab - 点击feed流banner",
-      "diary_item_click_comment": "日记本卡片 - 点击评论按钮",
-      "diary_item_click_image": "日记本卡片 - 点击图片",
-      "diary_item_click_vote": "日记本卡片 - 点赞",
+    "diary_item_click_avatar": "日记本卡片-点击作者头像",
+    "diary_item_click_banner": "首页精选tab-点击feed流banner",
+    "diary_item_click_comment": "日记本卡片-点击评论按钮",
+    "diary_item_click_image": "日记本卡片-点击图片",
+    "diary_item_click_vote": "日记本卡片-点赞",
     "diary_list_click_item": "跳转到日记本详情页",
     "diary_order_wow_click": "订单评价弹窗_点击",
-      "diary_topic_image_click": "日记贴创建页 - 点击已添加的图片",
+    "diary_topic_image_click": "日记贴创建页-点击已添加的图片",
     "diarybook_detail_bottom_click_collect": "日记详情页点击收藏按钮",
     "diarybook_detail_click_add_diary": "日记详情页点击更新日记按钮",
     "diarybook_detail_click_adjust": "日记详情页点击手术时间不准按钮",
@@ -241,7 +240,7 @@ name_dict = {
     "diarybook_detail_click_service_block": "日记详情页点击顶部美购卡片",
     "diarybook_detail_click_service_related": "日记详情页点击美购卡片",
     "diarybook_detail_click_sort": "日记详情页点击排序项按钮",
-      "diarybook_detail_click_tag_block": "点击顶部 TAG区域",
+    "diarybook_detail_click_tag_block": "点击顶部TAG区域",
     "diarybook_detail_click_topafter_picture": "日记详情页点击顶部afte图",
     "diarybook_detail_click_unfold_topic": "日记详情页点击展开全文按钮",
     "diarybook_detail_click_view_all_comments": "日记详情页点击查看全部评论按钮",
@@ -433,7 +432,7 @@ name_dict = {
     "on_click_tab": "tab切换点击",
     "on_click_top_card": "跳转到问题详情页或者回答详情页",
     "on_click_ugc_diary": "搜索结果页（综合tab）-点击日记本卡片",
-      "on_click_ugc_diary_comment": "日记本卡片 - 点击评论icon",
+    "on_click_ugc_diary_comment": "日记本卡片-点击评论icon",
     "on_click_ugc_diary_follow": "点击日记本关注按钮",
     "on_click_ugc_diary_like": "日记本卡片点赞",
     "on_click_ugc_diary_more": "点击查看更多按钮",
@@ -643,14 +642,14 @@ name_dict = {
     "welfare_special_click_diary": "美购专题页点击精选案例栏的日记本卡片",
     "welfare_special_click_doctor_rank_item": "美购专题页点击医生排行榜",
     "welfare_special_click_expert": "点击医生列表中的医生",
-      "welfare_special_click_more": "点击外显美购专题的 banner 或 “查看更多” 按钮",
+    "welfare_special_click_more": "点击外显美购专题的banner或“查看更多”按钮",
     "welfare_special_click_related_service": "点击外显美购专题的美购",
     "welfare_special_click_service": "美购专题页点击美购卡片",
     "welfare_special_click_service_rank_item": "美购专题页点击排行榜中的美购卡片",
     "welfare_special_click_storey_tab": "美购专题页点击TAB按钮",
     "welfare_special_click_storey_template": "美购专题页点击图片",
     "welfare_special_click_tab": "美购专题页点击推荐内容栏中的卡片",
-      "welfare_special_click_tag": "点击美购分组 tag 按钮",
+    "welfare_special_click_tag": "点击美购分组tag按钮",
     "welfare_special_click_template": "美购专题页点击固定模版",
     "wiki_detail_click_adv": "点击广告位",
     "wiki_detail_click_benefits_item": "百科详情页点击推荐美购卡片",
@@ -937,7 +936,7 @@ for t in range(1, task_days):
     on t2.device_id=dev.device_id
 WHERE spam_pv.device_id IS NULL
   and dev.device_id is null
-    """.format(today_str=today_str,yesterday_str_format=yesterday_str_format,today_str_format=today_str_format)
+    """.format(today_str=today_str, yesterday_str_format=yesterday_str_format, today_str_format=today_str_format)
 
     print(new_urser_device_id_sql)
     new_urser_device_id_df = spark.sql(new_urser_device_id_sql)
@@ -945,10 +944,9 @@ WHERE spam_pv.device_id IS NULL
     new_urser_device_id_df.show(1)
     sql_res = new_urser_device_id_df.collect()
 
-
     print("-------------------------------")
-    for count,res in enumerate(sql_res):
-        print(count,res)
+    for count, res in enumerate(sql_res):
+        print(count, res)
 
     res_dict = {}
     maidian_sql = """select t1.device_id,t2.track as track from 
@@ -977,6 +975,7 @@ on t1.device_id = t2.cl_id""".format(partition_date=yesterday_str)
                 res_dict[one_key_word] = 1
     print(res_dict)
 
+    exception_list = ("report_status","is_open_push")
     for action in res_dict:
         partition_date = yesterday_str
         pid = hashlib.md5((partition_date + action).encode("utf8")).hexdigest()
@@ -984,9 +983,12 @@ on t1.device_id = t2.cl_id""".format(partition_date=yesterday_str)
         action_ch = name_dict.get(action)
         if not action_ch:
             action_ch = action
+        if action_ch in exception_list:
+            continue
         instert_sql = """replace into new_user_click_count(
                     partition_day,pid,action_count,action_ch,action) VALUES('{partition_day}','{pid}',{action_count},'{action_ch}','{action}');""".format(
-            partition_date=partition_date,partition_day=yesterday_str,pid=pid,action_count=action_count,action_ch=action_ch,action=action
+            partition_date=partition_date, partition_day=yesterday_str, pid=pid, action_count=action_count,
+            action_ch=action_ch, action=action
         )
         print(instert_sql)
         # cursor.execute("set names 'UTF8'")
@@ -998,5 +1000,3 @@ on t1.device_id = t2.cl_id""".format(partition_date=yesterday_str)
         print(res)
     # cursor.executemany()
     db.close()
-
-