Multiple foreign keys in one column? - php

Question

I am trying to create a database for a homepage for dogs.
I use mysql and php. Every dog has a dog_id (PRIMARY KEY), which it can be identified by.
Now, I would like to be able to link a dog to its siblings. The problem here is that the number of siblings might be anywhere between 0 to let's say 25. Therefore, I am unsure how to do this.
Here is my 1st attempt
TABLE `dogs` (
`dog_id` INT(11) unsigned NOT NULL AUTO_INCREMENT COMMENT 'auto incremented PRIMARY KEY',
`dog_sibling1` INT(11) COLLATE utf8_unicode_ci NULL COMMENT 'FOREIGN KEY dog_id of the dog\'s 1st sibling',
`dog_sibling2` INT(11) COLLATE utf8_unicode_ci NULL COMMENT 'FOREIGN KEY dog_id of the dog\'s 2nd sibling',
and so on....
);
The problem I see here is that worst case I end up with 25 columns of NULL and most of the time with a lot of NULL columns.
My 2nd attempt
TABLE `dogs` (
`dog_id` INT(11) unsigned NOT NULL AUTO_INCREMENT COMMENT 'auto incremented PRIMARY KEY',
`dog_sibling_ids` INT(11) COLLATE utf8_unicode_ci NULL COMMENT 'FOREIGN KEY dog_ids of the dog\'s siblings',
);
Like this I would just have one column but I do not know if it is even possible to save an array of foreign keys in one column? And I also would not be sure how to deal with this in php then.
Is there a better way to solve this problem?

Answer 1

Your best bet would likely be to have a second table that 'groups' the siblings together.
CREATE TABLE dog_siblings (
sibling_group_id INTEGER,
dog_id integer
);
This way, if dog_ids 5-9 are all related, they just need to share a common sibling_group_id in the new table, and you can find all dogs related using a query similar to the following
SELECT ds2.dog_id
FROM dogs d
INNER JOIN dog_siblings ds
ON d.dog_id = ds.dog_id AND d.dog_id = 1
INNER JOIN dog_siblings ds2
ON ds.sibling_group_id = ds2.sibling_group_id
And that'll give you all dogs who are siblings of the dog with id 1 (including dog 1)
(edit: stupid inability to comment, the comments on the op appeared while i was crafting this, honest)

Answer 2

I would make a solution similar to the one you suggest in your question: create one table called siblings and reference it from dogs table:
CREATE TABLE `siblings` (
`siblings_id` INTEGER PRIMARY KEY
);
CREATE TABLE `dogs` (
`dog_id` INTEGER PRIMARY KEY,
`dog_sibling_id` INTEGER REFERENCES `siblings`
);
I wouldn't put any reference from the siblings table to the dogs table as it is clearly a one-to-many relationship (one siblings group can have many dogs but a dog can only belong to one siblings group). What about the dogs that have no siblings? You can either create a siblings group for those dogs also or just leave dog_sibling_id as NULL.
The previous works if you want to know the siblings of one litter (group of baby dogs born at the same time). If you need to track the siblings between different litters add the mother and father references:
CREATE TABLE `litters` (
`litter_id` INTEGER PRIMARY KEY,
`mother_id` INTEGER REFERENCES `dogs`,
`father_id` INTEGER REFERENCES `dogs`
);
CREATE TABLE `dogs` (
`dog_id` INTEGER PRIMARY KEY,
`dog_litter_id` INTEGER REFERENCES `litters`
);
You could also just add the mother and father columns to the dogs table and drop the litters table (in which case the litter specific information would be lost).