[英]Group by array values in Presto
有一個像這樣的電影表:
genre revenue
array<string> int
---------------------------
[Drama] 10000
[Drama, Fiction] 2000
[Fiction] 300
[Comedy] 5000
[Comedy, Fiction] 500
如何按照包含的流派對電影進行分組並執行聚合,例如:
genre count_movies sum_revenue
---------------------------------------
Drama 2 12000
Fiction 3 2800
Comedy 2 5500
我查看了數組函數和運算符的文檔,但在那個方向找不到任何東西。
潛在地,負責此輸出的查詢看起來像:
SELECT
COUNT(*) AS count_movies,
SUM(revenue) AS sum_revenue
FROM movies
GROUP BY ARRAY_EXPLODE(genre) -- this is fake, just to illustrate the point
你可以使用UNNEST
:
-- pseudocode
SELECT t.genre,
COUNT(*) AS count_movies,
SUM(revenue) AS sum_revenue
FROM movies
CROSS JOIN UNNEST(genre) AS t(genre)
GROUP BY t.genre
聲明:本站的技術帖子網頁,遵循CC BY-SA 4.0協議,如果您需要轉載,請注明本站網址或者原文地址。任何問題請咨詢:yoyou2525@163.com.