在sqlite3

Question

我有下表：

CREATE TABLE Kittens (
  k_id INTEGER PRIMARY KEY AUTOINCREMENT,
  human_name text,
  tags json
);

使用此数据

INSERT INTO Kittens (human_name, tags) VALUES ("Charles",'["fiendish", "fluffy", "quiet", "discerning", "catlike"]');
INSERT INTO Kittens (human_name, tags) VALUES ("Tails",'["fluffy", "loud","catlike"]');
INSERT INTO Kittens (human_name, tags) VALUES ("Mittens",'["fluffy","regal","catlike"]');

我想聚合常见的小猫标签，我相信这需要json_group_array功能，我在这里读到了这个：

• https://www.sqlite.org/json1.html#jarray
• http://sqlite.1065341.n5.nabble.com/json-group-array-td87148.html

但是，这似乎并没有像我期望的那样展开或聚合标签，只是连续列出列表。

sqlite> select json_group_array(tags) from Kittens;
["[\"fiendish\", \"fluffy\", \"quiet\", \"discerning\", \"catlike\"]","[\"fluffy\", \"loud\",\"catlike\"]","[\"fluffy\",\"regal\",\"catlike\"]"]

有关此查询应该是什么样子的任何指导？寻求更像的东西：

tag_array

catlike
catlike
catlike
discerning
fiendish
fluffy
fluffy
fluffy   
loud 
quiet 
regal

最终，聚合：

tag_array  count(*)
fluffy     3
catlike    3
fiendish   1
discerning 1
loud       1
regal      1
quiet      1

Answer 1

json_group_array()从其元素构造一个数组;要从数组中提取值，请使用json_each()：

SELECT tag, count(*)
FROM (SELECT j.value AS tag
      FROM Kittens
      CROSS JOIN json_each(Kittens.tags) AS j)
GROUP BY tag;