Commit 4d3b8cd8 authored by litaolemo's avatar litaolemo

update

parent 10337cb5
# -*- coding:UTF-8 -*-
# @Time : 2020/9/14 14:52
# @File : __init__.py.py
# @email : litao@igengmei.com
# @author : litao
\ No newline at end of file
# -*- coding:UTF-8 -*-
# @Time : 2020/9/14 14:53
# @File : meigou_huidu_huisu.py
# @email : litao@igengmei.com
# @author : litao
# -*- coding:UTF-8 -*-
# @Time : 2020/9/4 17:07
# @File : search_meigou_ctr.py
# @email : litao@igengmei.com
# @author : litao
import hashlib
import json
import pymysql
import xlwt, datetime
import redis
# from pyhive import hive
from maintenance.func_send_email_with_file import send_file_email
from typing import Dict, List
from elasticsearch_7 import Elasticsearch
from elasticsearch_7.helpers import scan
import sys
import time
from pyspark import SparkConf
from pyspark.sql import SparkSession, DataFrame
# from pyspark.sql.functions import lit
# import pytispark.pytispark as pti
def con_sql(sql):
# 从数据库的表里获取数据
db = pymysql.connect(host='172.16.40.158', port=4000, user='st_user', passwd='aqpuBLYzEV7tML5RPsN1pntUzFy',
db='jerry_prod')
cursor = db.cursor()
cursor.execute(sql)
result = cursor.fetchall()
db.close()
return result
startTime = time.time()
sparkConf = SparkConf()
sparkConf.set("spark.sql.crossJoin.enabled", True)
sparkConf.set("spark.debug.maxToStringFields", "100")
sparkConf.set("spark.tispark.plan.allow_index_double_read", False)
sparkConf.set("spark.tispark.plan.allow_index_read", True)
sparkConf.set("spark.hive.mapred.supports.subdirectories", True)
sparkConf.set("spark.hadoop.mapreduce.input.fileinputformat.input.dir.recursive", True)
sparkConf.set("spark.serializer", "org.apache.spark.serializer.KryoSerializer")
sparkConf.set("mapreduce.output.fileoutputformat.compress", False)
sparkConf.set("mapreduce.map.output.compress", False)
sparkConf.set("prod.gold.jdbcuri",
"jdbc:mysql://172.16.30.136/doris_prod?user=doris&password=o5gbA27hXHHm&rewriteBatchedStatements=true")
sparkConf.set("prod.mimas.jdbcuri",
"jdbc:mysql://172.16.30.138/mimas_prod?user=mimas&password=GJL3UJe1Ck9ggL6aKnZCq4cRvM&rewriteBatchedStatements=true")
sparkConf.set("prod.gaia.jdbcuri",
"jdbc:mysql://172.16.30.143/zhengxing?user=work&password=BJQaT9VzDcuPBqkd&rewriteBatchedStatements=true")
sparkConf.set("prod.tidb.jdbcuri",
"jdbc:mysql://172.16.40.158:4000/eagle?user=st_user&password=aqpuBLYzEV7tML5RPsN1pntUzFy&rewriteBatchedStatements=true")
sparkConf.set("prod.jerry.jdbcuri",
"jdbc:mysql://172.16.40.158:4000/jerry_prod?user=st_user&password=aqpuBLYzEV7tML5RPsN1pntUzFy&rewriteBatchedStatements=true")
sparkConf.set("prod.tispark.pd.addresses", "172.16.40.158:2379")
sparkConf.set("prod.tispark.pd.addresses", "172.16.40.170:4000")
sparkConf.set("prod.tidb.database", "jerry_prod")
sparkConf.setAppName("search_diary_ctr")
spark = (SparkSession.builder.config(conf=sparkConf).config("spark.sql.extensions", "org.apache.spark.sql.TiExtensions")
.config("spark.tispark.pd.addresses", "172.16.40.170:2379").appName(
"search_diary_ctr").enableHiveSupport().getOrCreate())
spark.sql("ADD JAR hdfs:///user/hive/share/lib/udf/brickhouse-0.7.1-SNAPSHOT.jar")
spark.sql("ADD JAR hdfs:///user/hive/share/lib/udf/hive-udf-1.0-SNAPSHOT.jar")
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 = 50
for t in range(1, task_days):
day_num = 0 - t
now = (datetime.datetime.now() + datetime.timedelta(days=day_num))
last_30_day_str = (now + datetime.timedelta(days=-30)).strftime("%Y%m%d")
today_str = now.strftime("%Y%m%d")
yesterday_str = (now + datetime.timedelta(days=-1)).strftime("%Y%m%d")
one_week_age_str = (now + datetime.timedelta(days=-7)).strftime("%Y%m%d")
sql_search_ctr = r"""
SELECT
t1.partition_date as `日期`
,active_type as `系统`
,device_os_type as `活跃`
,grey_type as `灰度类型`
,round(NVL(sum(click_pv),0)/NVL(sum(exp_pv),0)*100,2) as `卡片点击pv/卡片精准曝光pv(%)`
,round(NVL(sum(two_click_pv),0)/NVL(sum(exp_pv),0)*100,2) as `有效二跳pv/卡片精准曝光pv(%)`
,round(NVL(sum(two_click_pv),0)/NVL(sum(click_pv),0)*100,2) as `有效二跳pv/卡片点击pv(%)`
,round(NVL(sum(cpc_exp_pv),0)/NVL(sum(exp_pv),0)*100,2) as `cpc卡片曝光pv/卡片精准曝光pv(%)`
,NVL(sum(click_pv),0) as `卡片点击pv`
,NVL(sum(exp_pv),0) as `卡片曝光pv`
,NVL(sum(two_click_pv),0) as `有效二跳pv`
,NVL(sum(cpc_click_pv),0) as `cpc卡片点击pv`
,NVL(sum(cpc_exp_pv),0) as `cpc卡片曝光pv`
FROM
(
SELECT partition_date
,device_os_type
,CASE WHEN active_type = '4' THEN '老活'
WHEN active_type IN ('1','2') THEN '新增' END AS active_type
,device_id
,CASE WHEN substr(md5(device_id),-1) in ('0','1','2','3','4','5','6','7') THEN '灰度' ELSE '非灰' END AS grey_type
FROM online.ml_device_day_active_status
WHERE partition_date>={start_day}
AND partition_date<= {partition_day}
AND active_type IN ('1','2','4')
)t1
JOIN
(--精准曝光
SELECT cl_id,partition_date,card_id,count(1) as exp_pv,count(CASE WHEN get_json_object(exposure_card, '$.is_cpc')=1 THEN 1 END) as cpc_exp_pv
FROM online.ml_community_precise_exposure_detail
WHERE partition_date>={start_day}
AND partition_date<= {partition_day}
AND action in ('page_precise_exposure','home_choiceness_card_exposure') --7745版本action改为page_precise_exposure
AND page_name in('welfare_home')
AND tab_name in ('精选')
AND card_content_type ='service'
and (get_json_object(exposure_card,'$.in_page_pos')='' or get_json_object(exposure_card,'$.in_page_pos') is null)
group by partition_date,cl_id,card_id
)t2
on t1.device_id=t2.cl_id and t1.partition_date=t2.partition_date
LEFT JOIN
(--卡片点击
SELECT cl_id,partition_date,params['card_id'] as card_id,count(1) as click_pv,count(CASE WHEN params['is_cpc']=1 THEN 1 ELSE 0 END) as cpc_click_pv
FROM online.bl_hdfs_maidian_updates
WHERE partition_date>={start_day}
AND partition_date<= {partition_day}
AND action='on_click_card'
AND params['tab_name']='精选'
AND params['page_name'] ='welfare_home'
AND params['card_content_type'] ='service'
GROUP BY cl_id,partition_date,params['card_id']
)t3
on t2.partition_date=t3.partition_date
and t2.cl_id=t3.cl_id
and t2.card_id=t3.card_id
LEFT JOIN
(--商祥二跳
SELECT cl_id,partition_date,params['service_id'] as service_id,count(1) as two_click_pv
FROM online.bl_hdfs_maidian_updates
WHERE partition_date>={start_day}
AND partition_date<= {partition_day}
AND (referrer in ('welfare_home')
or (params['referrer_link'] like '%[%' and json_split(params['referrer_link'])[size(json_split(params['referrer_link']))-1] in ('welfare_home')))
AND ((action in ('welfare_multiattribute_click_add','welfare_multiattribute_click_buy') AND page_name = 'welfare_detail')
or action = 'welfare_detail_click_message')
GROUP BY cl_id,partition_date,params['service_id']
)t4
on t3.partition_date=t4.partition_date
and t3.cl_id=t4.cl_id
and t3.card_id=t4.service_id
LEFT JOIN
(
SELECT distinct device_id
FROM dim.dim_device_user_staff --去除内网用户
UNION ALL
SELECT device_id
FROM ml.ml_d_ct_dv_devicespam_d --剔除刷量设备
WHERE partition_day={partition_day}
)a
on t1.device_id=a.device_id
LEFT JOIN
(
SELECT partition_date,device_id
FROM
(--找出user_id当天活跃的第一个设备id
SELECT user_id,partition_date,
if(size(device_list) > 0, device_list [ 0 ], '') AS device_id
FROM online.ml_user_updates
WHERE partition_date>={start_day}
AND partition_date<= {partition_day}
)t1
JOIN
( --医生账号
SELECT distinct user_id
FROM online.tl_hdfs_doctor_view
WHERE partition_date = {partition_day}
--马甲账号/模特用户
UNION ALL
SELECT user_id
FROM ml.ml_c_ct_ui_user_dimen_d
WHERE partition_day = {partition_day}
AND (is_puppet = 'true' or is_classifyuser = 'true')
UNION ALL
--公司内网覆盖用户
select distinct user_id
from dim.dim_device_user_staff
UNION ALL
--登陆过医生设备
SELECT distinct t1.user_id
FROM
(
SELECT user_id, v.device_id as device_id
FROM online.ml_user_history_detail
LATERAL VIEW EXPLODE(device_history_list) v AS device_id
WHERE partition_date = {partition_day}
)t1
JOIN
(
SELECT device_id
FROM online.ml_device_history_detail
WHERE partition_date = {partition_day}
AND is_login_doctor = '1'
)t2
ON t1.device_id = t2.device_id
)t2
on t1.user_id=t2.user_id
group by partition_date,device_id
)b
on t1.partition_date=b.partition_date and t1.device_id=b.device_id
where (a.device_id is NULL or a.device_id ='')
and (b.device_id is null or b.device_id ='')
GROUP BY t1.partition_date
,grey_type,active_type,device_os_type
order by 1
""".format(partition_day=today_str,start_day=yesterday_str)
print(sql_search_ctr)
search_ctr_df = spark.sql(sql_search_ctr)
# spam_pv_df.createOrReplaceTempView("dev_view")
search_ctr_df.show(1)
sql_res = search_ctr_df.collect()
for res in sql_res:
print(res)
task_list.append(res)
print(task_list)
# cursor.executemany()
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