雪花中的值

Question

我有一张这样的桌子——

n0   n1  n2
232  12  211
323  45  5655

我想产生这个输出 -

n    Count
232  1
12   1 
211  1
323  1
45   1
5655 1

我可以在 PostgreSQL 中使用这个查询来做到这一点。

select el, count(*)
from 
(
SELECT  n0, n1, n2 from npi
) n
cross join lateral
     (select v.el
      from (values (n0), (n1), (n2)) v(el)
     ) v
group by el

但是当我尝试在 Snowflake 中运行相同的查询时，出现此错误 -

Invalid expression [CORRELATION(N.N0)] in VALUES clause

Answer 1

使用UNION ALL替代形式：

WITH cte(el) AS (
  SELECT n0 FROM npi
  UNION ALL SELECT n1 FROM npi
  UNION ALL SELECT n2 FROM npi
)
SELECT el, COUNT(*)
FROM cte
GROUP BY el;

输出：

---

编辑：

使用 FLATTEN 与 ARRAY_CONSTRUCT 结合的版本：

SELECT value AS el, COUNT(*)
FROM(SELECT  n0, n1, n2 from npi) AS n
,TABLE(FLATTEN(input => array_construct(n.n0, n.n1, n.n2)))
GROUP BY value;

Answer 2

以下是使用 UNPIVOT 的版本：

create or replace table test (n0 int, n1 int, n2 int);

insert into test values (232, 12, 211), (323, 45, 5655);

select value, count(1) from test
unpivot(value for col_name in (n0, n1, n2))
group by value;

+-------+----------+
| VALUE | COUNT(1) |
|-------+----------|
|   232 |        1 |
|    12 |        1 |
|   211 |        1 |
|   323 |        1 |
|    45 |        1 |
|  5655 |        1 |
+-------+----------+