Aggregating lists stored as json in sqlite3

Question

I have the following table:

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

With this data

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"]');

I'd like to aggregate the common kitten tags, and I believe this requires the json_group_array function, which I've read about here:

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

However, this doesn't seem to be unrolling or aggregating the tags as I would expect, just list the lists consecutively.

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

Any guidance on what this query should look like? Seeking something more like:

tag_array

catlike
catlike
catlike
discerning
fiendish
fluffy
fluffy
fluffy   
loud 
quiet 
regal

And ultimately, with an aggregation:

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

Answer 1

json_group_array() constructs an array from its elements; to extract values from an array, use json_each() :

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