Split single VARCHAR field with many values, and insert them separately - mysql

Question

I'm migrating an old database data to a new one, and they used to store telephone numbers in the following format:
Example 1:
41.9044-9082;41.9044-9661;41.9851-9862;41.9984-0393;41.3399-9169;41.3997-7999;
Example 2:
41.3369-0102;41.8928-5992;
No telephones(empty):
;
How can I split these single VARCHAR fields with many values, and insert them separately?
Table example of how it is:
|#id_tel#|### number ####|#|client_id|#|
|# 1 #|111163;3554353;|#| 2 |#|
|# 2 #|222222; |#| 3 |#|
|# 3 #|; |#| 4 |#|
Table example of how I would like it to be:
|#id_tel#|### number ####|#|client_id|#|
|# 1 #|111163 |#| 2 |#|
|# 2 #|3554353 |#| 2 |#|
|# 3 #|222222 |#| 3 |#|

Answer 1

You could do that with nested calls of SUBSTRING_INDEX() and a numbers table. In my example I create the numbers table on the fly for up to 100 numbers.
Assuming a table old_tel with following CREATE TABLE statement:
CREATE TABLE old_tel (
id_tel INT,
`number` VARCHAR(200),
client_id INT
);
you get the splitted numbers with the client_id with this query:
SELECT
SUBSTRING_INDEX(SUBSTRING_INDEX(`number`, ';', n.n), ';', -1) value,
client_id
FROM
old_tel
CROSS JOIN (
SELECT
1 + a.N + b.N * 10 AS n
FROM
(SELECT 0 AS N UNION ALL SELECT 1 UNION ALL SELECT 2 UNION ALL SELECT 3 UNION ALL SELECT 4 UNION ALL SELECT 5 UNION ALL SELECT 6 UNION ALL SELECT 7 UNION ALL SELECT 8 UNION ALL SELECT 9) a
,(SELECT 0 AS N UNION ALL SELECT 1 UNION ALL SELECT 2 UNION ALL SELECT 3 UNION ALL SELECT 4 UNION ALL SELECT 5 UNION ALL SELECT 6 UNION ALL SELECT 7 UNION ALL SELECT 8 UNION ALL SELECT 9) b
ORDER BY n
) n
WHERE
n.n <= LENGTH(`number`) - LENGTH(REPLACE(`number`, ';', ''))
AND
SUBSTRING_INDEX(SUBSTRING_INDEX(`number`, ';', n.n), ';', -1) <> ''
ORDER BY
client_id, SUBSTRING_INDEX(SUBSTRING_INDEX(`number`, ';', n.n), ';', -1);
Assuming your new telephone numbers table looks nearly the same:
CREATE TABLE new_tel (
id_tel INT NOT NULL AUTO_INCREMENT PRIMARY KEY,
`number` VARCHAR(200),
client_id INT
);
you can fill this table with this simple INSERT statement using the first query:
INSERT INTO new_tel (`number`, client_id)
SELECT
SUBSTRING_INDEX(SUBSTRING_INDEX(`number`, ';', n.n), ';', -1) value,
client_id
FROM
old_tel
CROSS JOIN (
SELECT
1 + a.N + b.N * 10 AS n
FROM
(SELECT 0 AS N UNION ALL SELECT 1 UNION ALL SELECT 2 UNION ALL SELECT 3 UNION ALL SELECT 4 UNION ALL SELECT 5 UNION ALL SELECT 6 UNION ALL SELECT 7 UNION ALL SELECT 8 UNION ALL SELECT 9) a
,(SELECT 0 AS N UNION ALL SELECT 1 UNION ALL SELECT 2 UNION ALL SELECT 3 UNION ALL SELECT 4 UNION ALL SELECT 5 UNION ALL SELECT 6 UNION ALL SELECT 7 UNION ALL SELECT 8 UNION ALL SELECT 9) b
ORDER BY n
) n
WHERE
n.n <= LENGTH(`number`) - LENGTH(REPLACE(`number`, ';', ''))
AND
SUBSTRING_INDEX(SUBSTRING_INDEX(`number`, ';', n.n), ';', -1) <> ''
ORDER BY
client_id, SUBSTRING_INDEX(SUBSTRING_INDEX(`number`, ';', n.n), ';', -1);
Explanation
The inner subselect with the UNION ALL creates on the fly a numbers table. We restrict this on the number of substrings in the number column and filter out empty values.
SUBSTRING_INDEX(SUBSTRING_INDEX(`number`, ';', n.n), ';', -1) value
cuts out the n-th number that is separated by semicolon.
See it working in this Demo
**Note: ** This is very fast by avoiding agonizing row by row inserts.