clickhouse做漏斗分析

漏斗分析代码记录

SELECT media_source AS media_source, funnel_sort AS funnel_sort, count(1) AS total_count
FROM (SELECT media_source,arraySort(x ->-x, groupArray(count)) AS funnel_sortFROM (SELECT media_source, level_index, count(1) AS countFROM (SELECT media_source,user_id                           AS user_id,arrayJoin(arrayEnumerate(levels)) AS level_index,arrayWithConstant(level, 1)       AS levelsFROM (SELECT media_source,user_id,windowFunnel(86400)(event_time, event_name = 'bf_enter_success',event_name = 'RewardVideoShowSuccess') AS levelFROM (SELECT date                                                            AS date,country                                                         AS country,app_name                                                        AS app_name,app_version                                                     AS app_version,media_source                                                    AS media_source,user_id                                                         AS user_id,event_name                                                      AS event_name,app_id                                                          AS app_id,toUInt64(toUnixTimestamp64Milli(toDateTime64(splitByString('+', event_time)[1], 3,'Asia/Shanghai'))) AS event_timeFROM ods.ods_appsflyer_eventWHERE 1 = 1AND date BETWEEN '2020-10-10' AND '2020-10-10'AND app_id = ('APAFXD4SDC4WE'))WHERE 1 = 1GROUP BY media_source, user_id)WHERE 1 = 1)WHERE 1 = 1GROUP BY media_source, level_index)WHERE 1 = 1GROUP BY media_source)
WHERE 1 = 1
GROUP BY media_source, funnel_sort
ORDER BY media_source ASC;

SELECT date AS date, count(user_id) AS count, SUM(data_0) AS data_0, SUM(data_1) AS data_1
FROM (SELECT date                                         AS date,user_id                                      AS user_id,groupArray(event)                            AS events,toString(events)                             AS events_uniq_str,arrayCount(x-> x = 1, events)                AS data_0,length(extractAll(events_uniq_str, '1.*?2')) AS data_1FROM (SELECT date                                                                          AS date,country                                                                       AS country,app_name                                                                      AS app_name,app_version                                                                   AS app_version,media_source                                                                  AS media_source,user_id                                                                       AS user_id,casewhen event_name = 'bf_enter_success' then 1when event_name = 'RewardVideoShowSuccess' then 2else 0 end                                                                AS event,app_id                                                                        AS app_id,toUInt64(toUnixTimestamp64Milli(toDateTime64(splitByString('+', event_time)[1], 3, 'Asia/Shanghai'))) AS event_timeFROM ods.ods_appsflyer_eventWHERE 1 = 1AND date BETWEEN '2020-10-10' AND '2020-10-10'AND app_id = ('APAFXD4SDC4WE')AND event_name in ('bf_enter_success', 'RewardVideoShowSuccess')ORDER BY event_time ASC)WHERE 1 = 1GROUP BY date, user_id)
WHERE 1 = 1
GROUP BY date
ORDER BY date ASC

clickhouse做漏斗分析相关推荐

在 ClickHouse 中使用函数实现有序漏斗分析
文章目录背景基础提出问题分析问题解决问题总结背景最近大数据项目中,碰到了个问题,在做漏斗分析时分析性能常常跟不上,22 亿数据量往往需要 10s 以上才能返回想要的结果.推测应该是分析 ...
ClickHouse 实现有序漏斗分析与数据可视化
Clickhouse 实现漏斗分析与数据可视化 1.前言 2.环境准备 2.1.测试表 2.2.插入伪造的样例数据 2.2.1.插入伪造的登录数据 2.2.2.插入伪造的浏览数据 2.2.3.插入伪造 ...
如何简单、高效的进行留存和漏斗分析
Databend 是一个使用 Rust 研发.开源的.完全面向云架构的新式数仓,致力于提供极速的弹性扩展能力,打造按需.按量的 Data Cloud 产品体验. 开源地址:https://github ...
九大数据分析方法：漏斗分析法
大家好,我是爱学习的小xiong熊妹. 今天继续跟大家分享的是九大数据分析方法系列.今天介绍的是漏斗分析法,漏斗分析法是一种基础的,处理多个指标分析问题的方法,有很多应用场景. 一.为什么叫" ...
一文读懂大数据漏斗分析！
下面我将为您介绍漏斗分析的概念,以及组织及其数据如何使用它来分析大数据集.希望能够帮到你们. 很多时候我们发现我们的业务数据看起来很不错,但最终的转换量却很低,这就说明业务转换率这里有问题.而在数 ...
神策数据罗彦博：如何正确使用漏斗分析提升转化？
本文根据神策数据数字营销经理罗彦博<如何正确使用漏斗分析提升转化>直播整理而成.本文主要内容如下: 通过漏斗分析做 SEM 实时调整推广落地页的转化路径设计 APP 推广中的数据拦截问题 ...
aspen怎么做灵敏度分析_不会敲代码怎么做智能分析？用奥威商业智能BI
不会敲代码,怎么学也不会,那怎么做大数据智能分析?说来也简单,用商业智能BI软件就行.拖拉拽就能做报表,点击即可智能分析,即使是没有IT基础,完全不会敲代码的人也能轻松抽取数据做智能分析. 商业智能B ...
数据分析模型：漏斗分析
01 什么是漏斗分析漏斗分析是一套流程式的数据分析方法,能够科学地反映各阶段用户转化情况. 漏斗分析模型已经广泛应用于用户行为分析类产品,且功能十分强大:它可以评估总体或各个环节的转化情况.促销活动 ...
承上启下：基于全域漏斗分析的主搜深度统一粗排
1. 背景 1.1 概述淘宝主搜索是一个典型的多阶段检索系统,主要分为召回.粗排.精排等阶段.召回阶段,由文本召回.个性化等多路召回构成,输出商品量级约10^5:粗排阶段,需要从三路召回集合中分别进 ...

clickhouse做漏斗分析

clickhouse做漏斗分析相关推荐

最新文章

热门文章