SQL join on two table and replace and return columns

Question

I am working on some SQL query with joins and I want to get the corresponding value from another table which matches with original table columns values.

For example, I have two different tables named Product and Category and they are as follows:

Products:

| id | name    | category1                 | category2    | category3   |
| -- | ------- | ------------------------- |--------------------------- |
| 1  | T-shirt | Cloths&Accessories~:/ID1  | Cloths~:/ID2 | Shirts~:/ID3|
|----|---------|---------------------------|------------- |-------------|
| 2  | T-shirt | Cloths&Accessories~:/ID1  | Cloths~:/ID2 | Shirts~:/ID3|
|----|---------|---------------------------|------------- |-------------|
| 3  | T-shirt | Cloths&Accessories~:/ID1  | Cloths~:/ID2 | Shirts~:/ID3|
|----|---------|---------------------------|------------- |-------------|
| 4  | T-shirt | Cloths&Accessories~:/ID1  | Cloths~:/ID2 | Shirts~:/ID3|
|----|---------|---------------------------|------------- |-------------|
| 5  | T-shirt | Cloths&Accessories~:/ID1  | Cloths~:/ID2 | Shirts~:/ID3|
|----|---------|---------------------------|------------- |-------------|

And Category:

| id | categories           |
| -- | -------------------- |
| 1  | Cloths & Accessories |
|----|--------------------- |
| 2  | Cloths               |
|----|--------------------- |
| 3  | Shirts               |
|----|--------------------- |

The problem is, in table Product there is this Cloths&Accessories~:/ID1 kind of Invalid strings stored under column names category1 , category2 and category3 and I want to replace them with valid category names from table Category . If you look at the table Category the id of each category matches with string Cloths&Accessories~:/ID1 having ID1 in it.

For example: Cloths&Accessories~:/ID1 in this string ID1 is related to id=1 of table Category which is Cloths & Accessories

I have to replace and return all of category1 , category2 and category3 columns of table Product with valid categories .

What would be the optimal SQL join query for this?

Answer 1

You could use substring_index to extract the ID, and then join on it:

SELECT p.id AS id,
       p.name AS name,
       c1.categories AS category1,
       c2.categories AS category2,
       c3.categories AS category3
FROM   products p
JOIN   category c1 ON SUBSTRING_INDEX(p.category1, 'ID', -1) = c1.id
JOIN   category c2 ON SUBSTRING_INDEX(p.category2, 'ID', -1) = c2.id
JOIN   category c3 ON SUBSTRING_INDEX(p.category3, 'ID', -1) = c3.id