▲ | gigatexal a day ago | |||||||
In a well normalized setup idk maybe not. Uuidv4 for your external ids and then have a mapping table to correspond that to something you’d use internally. Then you can torch an exposed uuid update the mapping table and generate a new one and none of your pointers and foreign keys need to change internally. | ||||||||
▲ | crazygringo a day ago | parent | next [-] | |||||||
The point is, that mapping table incurs the same indexing cost that was trying to be eliminated in the first place. Normalization is irrelevant. | ||||||||
| ||||||||
▲ | Quekid5 a day ago | parent | prev [-] | |||||||
I wonder if there is a name for such a mapping table in RDBMS-land...? | ||||||||
|