尝试替换雪花上的空值时使用 LEAD 和 LAG 时出现不一致

Question

我是 snowflake 的新手，正在尝试编写一个 SQL 查询，以根据日期为每个 ID 将空值替换为最后记录的 Ip。 Id被认为是降序的，每个Id的日期也是降序的。

这是我的例子table

table姓名：客户

id	date	ip
113	2022-02-05 11:40:42	Null
113	2021-12-05 11:40:42	Null
113	2021-08-05 11:40:42	Null
113	2021-07-05 11:40:42	Null
113	2022-02-05 11:40:42	83:93:225:63
112	2022-02-05 11:40:42	Null
112	2021-02-05 11:40:42	3:9:225:63
112	2020-02-05 11:40:42	8:9:225:63

我想达到的目标

id	date	ip
113	2022-02-05 11:40:42	83:93:225:63
113	2021-12-05 11:40:42	83:93:225:63
113	2021-08-05 11:40:42	83:93:225:63
113	2021-07-05 11:40:42	83:93:225:63
113	2022-02-05 11:40:42	83:93:225:63
112	2022-02-05 11:40:42	3:9:225:63
112	2021-02-05 11:40:42	3:9:225:63
112	2020-02-05 11:40:42	8:9:225:63

Answer 1

我不确定你是否想要“下一个”、“最后一个”、“前一个”值，所以这三个都在这里

已编辑为与您要求的排序顺序相同：

SELECT column1 as id
    ,to_date(column2) as date
    ,column3 as ip 
    ,LAST_VALUE(ip) IGNORE NULLS OVER (PARTITION BY id ORDER BY date) last_value
    ,NVL(ip, last_value) AS current_or_last
    ,LEAD(ip) IGNORE NULLS OVER (PARTITION BY id ORDER BY date) as next_value
    ,NVL(ip, next_value) AS current_or_next  
    ,LAG(ip) IGNORE NULLS OVER (PARTITION BY id ORDER BY date) as prior_value
    ,NVL(ip, prior_value) AS current_or_prior
FROM VALUES
    (113,'2022-02-05 11:40:42',null),
    (113,'2021-12-05 11:40:42',null),
    (113,'2021-08-05 11:40:42',null),
    (113,'2021-07-05 11:40:42',null),
    (113,'2022-02-05 11:40:42','83:93:225:63'),
    (112,'2022-02-05 11:40:42',null),
    (112,'2021-02-05 11:40:42','3:9:225:63'),
    (112,'2020-02-05 11:40:42','8:9:225:63')
ORDER BY 1 DESC,2 DESC;

给出：

ID	DATE	IP	LAST_VALUE	CURRENT_OR_LAST	NEXT_VALUE	CURRENT_OR_NEXT	PRIOR_VALUE	CURRENT_OR_PRIOR
113	2022-02-05		83:93:225:63	83:93:225:63	83:93:225:63	83:93:225:63
113	2022-02-05	83:93:225:63	83:93:225:63	83:93:225:63		83:93:225:63		83:93:225:63
113	2021-12-05		83:93:225:63	83:93:225:63	83:93:225:63	83:93:225:63
113	2021-08-05		83:93:225:63	83:93:225:63	83:93:225:63	83:93:225:63
113	2021-07-05		83:93:225:63	83:93:225:63	83:93:225:63	83:93:225:63
112	2022-02-05		3:9:225:63	3:9:225:63			3:9:225:63	3:9:225:63
112	2021-02-05	3:9:225:63	3:9:225:63	3:9:225:63		3:9:225:63	8:9:225:63	3:9:225:63
112	2020-02-05	8:9:225:63	3:9:225:63	8:9:225:63	3:9:225:63	8:9:225:63		8:9:225:63

猜想您希望 current_or_last 像这样组合在一起：

SELECT column1 as id
    ,to_date(column2) as date
    ,NVL(column3, LAST_VALUE(column3) IGNORE NULLS OVER (PARTITION BY id ORDER BY date)) AS ip
FROM VALUES
    (113,'2022-02-05 11:40:42',null),
    (113,'2021-12-05 11:40:42',null),
    (113,'2021-08-05 11:40:42',null),
    (113,'2021-07-05 11:40:42',null),
    (113,'2022-02-05 11:40:42','83:93:225:63'),
    (112,'2022-02-05 11:40:42',null),
    (112,'2021-02-05 11:40:42','3:9:225:63'),
    (112,'2020-02-05 11:40:42','8:9:225:63')
ORDER BY 1 DESC,2 DESC;

给出：

ID	DATE	IP
113	2022-02-05	83:93:225:63
113	2022-02-05	83:93:225:63
113	2021-12-05	83:93:225:63
113	2021-08-05	83:93:225:63
113	2021-07-05	83:93:225:63
112	2022-02-05	3:9:225:63
112	2021-02-05	3:9:225:63
112	2020-02-05	8:9:225:63

尝试替换雪花上的空值时使用 LEAD 和 LAG 时出现不一致

inconsistencies while using LEAD and LAG when trying to replace null values on snowflake

sql

snowflake-cloud-data-platform