Why uuid() creates very similar identifiers in MariaDB?

I am creating a code generator for MariaDB to create a database based on a given JSON.

In that JSON, some initial data also exists. Thus I loop over data and insert them into the database.

Some columns have uuid() default value.

Here's the result of my code inserting data into such a table:

Id,Guid,Key,Order
1,5c52e1db-6809-11ec-982c-0242c0a81003,New,
2,5c530e55-6809-11ec-982c-0242c0a81003,WaitingForBusinessResponse,
3,5c533551-6809-11ec-982c-0242c0a81003,WaitingForUserResponse,
4,5c536433-6809-11ec-982c-0242c0a81003,UnderInvestigation,
5,5c538ba5-6809-11ec-982c-0242c0a81003,Closed,

As you can see UUID values are very very close to each other. This column has a unique index on it, so no duplicate entries would be allowed. But these values make it difficult to track them and one might easily confuse them with each other.

Is there a way to change this behavior? I want to tell MariaDB to create UUID more randomly.

Solution

The simple answer is "no, you can't" unless you write your own uuid function which provides an algorithm which creates a uuid() from random or pseudo random number as described in Chapter 4.4 of RFC 4122

The uuid() function of MariaDB (and MySQL) was implemented according to RFC 4122, but uses the algorithm for creating a time based uuid (see chapter 4.2)

Since all algorithms (name based, time based, random) deliver a Universal Unique Identifier which is globally unique in space and time, I don't really understand why you want to change the algorithm from time to random.

Time based uuids using uuidgen and mariadb:

~$ uuidgen -t;uuidgen -t
a5d3c032-6865-11ec-bd1f-1740cb8be951
a5d42d24-6865-11ec-bd1f-1740cb8be951
    
~$ mariadb -e"select uuid()\G";mariadb -e"select uuid()\G"
*************************** 1. row ***************************
uuid(): 45aca397-683c-11ec-a913-d83bbf89f2e2
*************************** 1. row ***************************
uuid(): 45ad94dd-683c-11ec-a913-d83bbf89f2e2