When comparing SQL that's generated using ULIDs, UUIDs, and other keys with stored test fixtures is difficult since each time the code is generated, the ULID values are not idempotent. UUIDv5 and similar algorithms create the same output for each input so perhaps we need a way to generate ULIDs in a "fixture comparison friendly way".
When comparing SQL that's generated using ULIDs, UUIDs, and other keys with stored test fixtures is difficult since each time the code is generated, the ULID values are not idempotent. UUIDv5 and similar algorithms create the same output for each input so perhaps we need a way to generate ULIDs in a "fixture comparison friendly way".