I’m curious about the actual format of the unique row IDs.
In one table, I get IDs like [25660,8955035] where all rows have the same prefix but the second part varies. The prefix doesn’t seem to be consistent between tables though.
Are the IDs assigned sequentially? does the first part reflect a unique database/table combination?
I’ve found a bit of info on the PostgreSQL CTID format. It looks like a tuple comprised of ( pageNumber , tupleNumber )
Here’s a link…
Back when Table ID numbers appeared in anvil.yaml
, I noticed that the 1st integer matched the table number.
However, that was not documented, and shouldn’t be relied on. For all we know, next week, a ton of row ids might get re-written behind the scenes, in some other format. Anvil reserves the right to rewrite internal IDs whenever it sees fit. As long as every copy of an ID, in the Anvil database, changes consistently, and at the same time, then it won’t matter to our Apps.
1 Like