How to collapse contiguous single day start/end date ranges in MySQL

Question

(edited - I oversimplified the original problem)

I'm trying to figure out a MySQL query that will collapse one-day long contiguous start/end date ranges. Here is my sample input data:

INSERT INTO my_table(tag, start_date, end_date) VALUES ('AA', '2020-07-01', '2020-07-02');
INSERT INTO my_table(tag, start_date, end_date) VALUES ('XB', '2020-07-02', '2020-07-03');
INSERT INTO my_table(tag, start_date, end_date) VALUES ('YC', '2020-07-03', '2020-07-04');
INSERT INTO my_table(tag, start_date, end_date) VALUES ('ZD', '2020-07-04', '2020-07-05');
INSERT INTO my_table(tag, start_date, end_date) VALUES ('0E', '2020-07-05', '2020-07-06');

INSERT INTO my_table(tag, start_date, end_date) VALUES ('CA', '2020-07-06', '2020-07-15');

INSERT INTO my_table(tag, start_date, end_date) VALUES ('CA', '2020-07-20', '2020-07-21');
INSERT INTO my_table(tag, start_date, end_date) VALUES ('CA', '2020-07-21', '2020-07-22');
INSERT INTO my_table(tag, start_date, end_date) VALUES ('CA', '2020-07-22', '2020-07-23');

INSERT INTO my_table(tag, start_date, end_date) VALUES ('EA', '2020-07-23', '2020-07-31');

Here is the output I want:

AA,2020-07-01,2020-07-06
CA,2020-07-06,2020-07-15
CA,2020-07-20,2020-07-23
EA,2020-07-23,2020-07-31

It seems like it should be simple, but I can't figure out how to do it in pure MySQL. I've tried joining the table to itself and using various combinations of MIN() and MAX(), but the solution eludes me. I can do it in PHP code, but would prefer to let MySQL do the work.

I can't rely on the tag field for grouping (it can be anything). The date collapse should be done only on the date ranges that are 1-day long and contiguous. Date ranges longer than one day should not be collapsed. The selected tag should be the value associated with the first date in the range

Answer 1

After a little more research, I figured out the solution to this "gaps and islands" problem:

SELECT t.tag, begin AS start_date, MAX(end) AS end_date
FROM (
    SELECT
    a.*
    CASE WHEN (DATEDIFF(start_date, @prev_date) = 1 AND DATEDIFF(end_date, start_date) = 1)
        THEN @begin := @begin
        ELSE @begin := start_date
    END AS begin,
    end_date AS end,
    @prev_date := start_date
    FROM (SELECT @prev_date := NULL, @begin := NULL) AS vars
    JOIN my_table a
    ORDER BY start_date
) AS t
GROUP BY begin;