给定 Snowflake 中的两个数组，找到最小元素的总和

Question

给定两个数组，如 a = [10, 20, 30] 和 b = [9, 21, 32]，我如何根据雪花中的索引构造一个由最小或最大元素组成的数组，即所需的最小输出是 [9,20,30]，最大值是 [10,21,32]?

我查看了 snowflake 的数组函数，但没有找到执行此操作的函数。

Answer 1

使用数字 table/[] 访问元素并使用 ARRAY_AGG 构建新数组：

WITH cte AS (
    SELECT ARRAY_CONSTRUCT(10, 20, 30) AS a, ARRAY_CONSTRUCT(9, 21, 32) AS b
), numbers AS (
  SELECT ROW_NUMBER() OVER(ORDER BY seq4())-1 AS IND 
  FROM TABLE(GENERATOR(ROWCOUNT => 10001))
)
SELECT a,b
    ,ARRAY_AGG(LEAST(a[ind], b[ind])) WITHIN GROUP(ORDER BY n.ind) AS min_array
    ,ARRAY_AGG(GREATEST(a[ind], b[ind]))  WITHIN GROUP(ORDER BY n.ind) AS max_array
FROM cte
JOIN numbers n
  ON n.ind < GREATEST(ARRAY_SIZE(a), ARRAY_SIZE(b))
GROUP BY a,b;

输出：

Answer 2

如果数组的大小始终相同（并重复使用 Lukasz 大数据 cte）：

WITH cte AS (
    SELECT ARRAY_CONSTRUCT(10, 20, 30) AS a, ARRAY_CONSTRUCT(9, 21, 32) AS b
)
SELECT a,b
    ,ARRAY_AGG(LEAST(a[n.index], b[n.index])) WITHIN GROUP(ORDER BY n.index) AS min_array
    ,ARRAY_AGG(GREATEST(a[n.index], b[n.index]))  WITHIN GROUP(ORDER BY n.index) AS max_array
FROM cte
    ,table(flatten(a)) n
GROUP BY 1,2;

给出：

A	B	MIN_ARRAY	MAX_ARRAY
[ 10, 20, 30 ]	[ 9, 21, 32 ]	[ 9, 20, 30 ]	[ 10, 21, 32 ]

如果您的列表不均匀：

WITH cte AS (
    SELECT ARRAY_CONSTRUCT(10, 20, 30) AS a, ARRAY_CONSTRUCT(9, 21, 32) AS b
    union all
    SELECT ARRAY_CONSTRUCT(10, 20, 30) AS a, ARRAY_CONSTRUCT(9, 21, 32, 45) AS b
)
SELECT a,b
    ,ARRAY_AGG(LEAST(a[n.index], b[n.index])) WITHIN GROUP(ORDER BY n.index) AS min_array
    ,ARRAY_AGG(GREATEST(a[n.index], b[n.index]))  WITHIN GROUP(ORDER BY n.index) AS max_array
FROM cte
    ,table(flatten(iff(array_size(a)>=array_size(b), a, b))) n
GROUP BY 1,2;

A	B	MIN_ARRAY	MAX_ARRAY
[ 10, 20, 30 ]	[ 9, 21, 32 ]	[ 9, 20, 30 ]	[ 10, 21, 32 ]
[ 10, 20, 30 ]	[ 9, 21, 32, 45 ]	[ 9, 20, 30 ]	[ 10, 21, 32 ]

会选择最大的，但是考虑到较小列表中的 NULL 会导致 LEAST/GREATEST 到 return NULL 并且 ARRAY_AGG 会丢弃空值，您甚至不需要进行大小比较，除非您想 NVL/COALESCE 将该值设置为空值的安全值。

SELECT 1 as a, null as b, least(a,b);

给出：

A	B	LEAST(A,B)
1	null	null

像这样：

SELECT a,b
    ,ARRAY_AGG(LEAST(nvl(a[n.index],10000), nvl(b[n.index],10000))) WITHIN GROUP(ORDER BY n.index) AS min_array
    ,ARRAY_AGG(GREATEST(nvl(a[n.index],0), nvl(b[n.index],0)))  WITHIN GROUP(ORDER BY n.index) AS max_array
FROM cte
    ,table(flatten(iff(array_size(a)>=array_size(b), a, b))) n
GROUP BY 1,2;

A	B	MIN_ARRAY	MAX_ARRAY
[ 10, 20, 30 ]	[ 9, 21, 32 ]	[ 9, 20, 30 ]	[ 10, 21, 32 ]
[ 10, 20, 30 ]	[ 9, 21, 32, 45 ]	[ 9, 20, 30, 45 ]	[ 10, 21, 32, 45 ]

给定 Snowflake 中的两个数组，找到最小元素的总和

Given two arrays in Snowflake, find summation of minimum elements

sql

arrays

snowflake-cloud-data-platform