update

task/spark_test.py

@@ -93,12 +93,12 @@ spark.sql("CREATE TEMPORARY FUNCTION arrayMerge AS 'com.gmei.hive.common.udf.UDF
 # huidu_device_id_df = spark.sql(huidu_device_id_sql)
 # huidu_device_id_df.createOrReplaceTempView("dev_view")
 sql_search_ctr = r"""
-select count(1),avg(session_pv),avg(session_pv2) from (SELECT  partition_date,
+ SELECT  partition_date,
+                    card_content_type,
                     cl_id,
                     v.recommend_type,
-                    count(distinct app_session_id) as session_pv ,
-                     count(app_session_id) as session_pv1,
-                     count(card_id) as session_pv2
+                    count(card_id) as session_pv0,
+                    count(distinct app_session_id) as session_pv
             FROM
             (
                SELECT partition_date,
@@ -123,7 +123,7 @@ select count(1),avg(session_pv),avg(session_pv2) from (SELECT  partition_date,
                       card_id,
                       app_session_id
                from online.ml_community_precise_exposure_detail
-               WHERE partition_date='20201105'
+               WHERE partition_date={partition_day}
                AND action in ('page_precise_exposure','home_choiceness_card_exposure') --7745版本action改为page_precise_exposure
                AND is_exposure = '1'  ----精准曝光
                AND page_name ='home'
@@ -153,8 +153,8 @@ select count(1),avg(session_pv),avg(session_pv2) from (SELECT  partition_date,
                       app_session_id
             )a
             LATERAL VIEW explode (a.recommend_type) v as recommend_type
-            group by partition_date,card_content_type,cl_id,v.recommend_type having  session_pv2 >0) 
-""".format(start_date='20201018',end_date='20201025')
+            group by partition_date,card_content_type,cl_id,v.recommend_type
+""".format(partition_day='20201105',end_date='20201025')
 
 print(sql_search_ctr)
 search_ctr_df = spark.sql(sql_search_ctr)