Skip to content

M
meta_base_code
Commit cf230ab4 authored by litaolemo's avatar litaolemo
Browse files
Options

update

parent a6956ac4
Hide whitespace changes
Inline Side-by-side
Showing with 0 additions and 165 deletions
task/search_strategy_d.py
View file @ cf230ab4
......@@ -175,90 +175,6 @@ for t in range(0, task_days):
)t
)t1
LEFT JOIN
(
SELECT t1.query,active_type,device_os_type,channel
,sum(hexin) as hexin_card_click_pv
,sum(neirong) as neirong_card_click_pv
FROM
(
SELECT NVL(t2.partition_date,t3.partition_date) as partition_date
,NVL(t2.cl_id,t3.cl_id) as cl_id
,NVL(t2.query,t3.query) as query
,NVL(t2.pv,0) as hexin
,NVL(t3.pv,0) as neirong
FROM
(
SELECT partition_date
,params['query'] as query
,cl_id
,count(1) as pv
FROM online.bl_hdfs_maidian_updates
WHERE partition_date >= {last_30_day_str}
AND partition_date < {today_str}
AND ((action in ('search_result_click_recommend_item','search_result_welfare_click_item','search_result_hospital_click_item','search_result_doctor_click_item','on_click_doctor_card', 'on_click_hospital_card')
AND page_name in ('search_result_more','search_result_welfare','search_result_hospital','search_result_doctor'))
or (action = 'goto_welfare_detail' AND params [ 'from' ] = 'search_result_welfare_recommend')
or (action = 'on_click_card' AND params['card_content_type'] in ('service','hospital','doctor') AND page_name in ('search_result_more','search_result_welfare','search_result_hospital','search_result_doctor'))
or (action = 'on_click_button' AND params['button_name'] = 'check_plan' AND page_name = 'search_result_more'))
GROUP BY partition_date
,params['query']
,cl_id
)t2
FULL JOIN
(
SELECT partition_date
,params['query'] as query
,cl_id
,count(1) as pv
FROM online.bl_hdfs_maidian_updates
WHERE partition_date >= {last_30_day_str}
AND partition_date < {today_str}
AND ((action in ('on_click_topic_card','on_click_diary_card','search_result_click_infomation_item')
AND page_name in ('search_result_more','search_result_diary','search_result_post'))
or (action = 'on_click_card' AND params['card_content_type'] in ('answer','diary') AND page_name in ('search_result_more','search_result_diary','search_result_question_answer')))
GROUP BY partition_date
,params['query']
,cl_id
)t3
on t3.partition_date=t2.partition_date
AND t3.query=t2.query
AND t3.cl_id=t2.cl_id
)t1
JOIN
(
SELECT partition_date,device_id,t2.active_type,t2.channel,t2.device_os_type
FROM
(
SELECT
partition_date,m.device_id
,array(device_os_type ,'合计') as device_os_type
,array(case WHEN active_type = '4' THEN '老活'
WHEN active_type in ('1','2') then '新增' END ,'合计') as active_type
,array(CASE WHEN is_ai_channel = 'true' THEN 'AI' ELSE '其他' END , '合计') as channel
FROM online.ml_device_day_active_status m
LEFT JOIN
(SELECT code,is_ai_channel,partition_day
FROM DIM.DIM_AI_CHANNEL_ZP_NEW
WHERE partition_day>= {last_30_day_str}
AND partition_day < {today_str}) tmp
ON m.partition_date=tmp.partition_day AND first_channel_source_type=code
WHERE partition_date >= {last_30_day_str}
AND partition_date < {today_str}
AND active_type in ('1','2','4')
) mas
LATERAL VIEW explode(mas.channel) t2 AS channel
LATERAL VIEW explode(mas.device_os_type) t2 AS device_os_type
LATERAL VIEW explode(mas.active_type) t2 AS active_type
)dev
on t1.cl_id=dev.device_id and t1.partition_date = dev.partition_date
GROUP BY t1.query,active_type,device_os_type,channel
)t2
on t1.query=t2.query and t1.active_type=t2.active_type and t1.device_os_type = t2.device_os_type AND t1.channel = t2.channel
LEFT JOIN
(
SELECT query,active_type,device_os_type,channel,search_pv,search_uv
......@@ -529,87 +445,6 @@ for t in range(0, task_days):
)t5
on t1.query=t5.query and t1.active_type=t5.active_type and t1.device_os_type = t5.device_os_type AND t1.channel = t5.channel
LEFT JOIN
(
SELECT t1.query,active_type,device_os_type,channel
,sum(hexin) as hexin_card_click_pv
,sum(neirong) as neirong_card_click_pv
FROM
(
SELECT NVL(t2.partition_date,t3.partition_date) as partition_date
,NVL(t2.cl_id,t3.cl_id) as cl_id
,NVL(t2.query,t3.query) as query
,NVL(t2.pv,0) as hexin
,NVL(t3.pv,0) as neirong
FROM
(
SELECT partition_date
,params['query'] as query
,cl_id
,count(1) as pv
FROM online.bl_hdfs_maidian_updates
WHERE partition_date >= {one_week_age_str}
AND partition_date < {today_str}
AND ((action in ('search_result_click_recommend_item','search_result_welfare_click_item','search_result_hospital_click_item','search_result_doctor_click_item','on_click_doctor_card', 'on_click_hospital_card')
AND page_name in ('search_result_more','search_result_welfare','search_result_hospital','search_result_doctor'))
or (action = 'goto_welfare_detail' AND params [ 'from' ] = 'search_result_welfare_recommend')
or (action = 'on_click_card' AND params['card_content_type'] in ('service','hospital','doctor') AND page_name in ('search_result_more','search_result_welfare','search_result_hospital','search_result_doctor'))
or (action = 'on_click_button' AND params['button_name'] = 'check_plan' AND page_name = 'search_result_more'))
GROUP BY partition_date
,params['query']
,cl_id
)t2
FULL JOIN
(
SELECT partition_date
,params['query'] as query
,cl_id
,count(1) as pv
FROM online.bl_hdfs_maidian_updates
WHERE partition_date >= {one_week_age_str}
AND partition_date < {today_str}
AND ((action in ('on_click_topic_card','on_click_diary_card','search_result_click_infomation_item')
AND page_name in ('search_result_more','search_result_diary','search_result_post'))
or (action = 'on_click_card' AND params['card_content_type'] in ('answer','diary') AND page_name in ('search_result_more','search_result_diary','search_result_question_answer')))
GROUP BY partition_date
,params['query']
,cl_id
)t3
on t3.partition_date=t2.partition_date
AND t3.query=t2.query
AND t3.cl_id=t2.cl_id
)t1
JOIN
(
SELECT partition_date,device_id,t2.active_type,t2.channel,t2.device_os_type
FROM
(
SELECT
partition_date,m.device_id
,array(device_os_type ,'合计') as device_os_type
,array(case WHEN active_type = '4' THEN '老活'
WHEN active_type in ('1','2') then '新增' END ,'合计') as active_type
,array(CASE WHEN is_ai_channel = 'true' THEN 'AI' ELSE '其他' END , '合计') as channel
FROM online.ml_device_day_active_status m
LEFT JOIN
(SELECT code,is_ai_channel,partition_day
FROM DIM.DIM_AI_CHANNEL_ZP_NEW
WHERE partition_day>= {one_week_age_str}
AND partition_day < {today_str}) tmp
ON m.partition_date=tmp.partition_day AND first_channel_source_type=code
WHERE partition_date >= {one_week_age_str}
AND partition_date < {today_str}
AND active_type in ('1','2','4')
) mas
LATERAL VIEW explode(mas.channel) t2 AS channel
LATERAL VIEW explode(mas.device_os_type) t2 AS device_os_type
LATERAL VIEW explode(mas.active_type) t2 AS active_type
)dev
on t1.cl_id=dev.device_id and t1.partition_date = dev.partition_date
GROUP BY t1.query,active_type,device_os_type,channel
)t6
on t1.query=t6.query and t1.active_type=t6.active_type and t1.device_os_type = t6.device_os_type AND t1.channel = t6.channel
""".format(last_30_day_str=last_30_day_str, today_str=today_str,yesterday_str=yesterday_str, one_week_age_str=one_week_age_str)
device_df = spark.sql(sql)
device_df.createOrReplaceTempView("data_table")
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment