|
|
|
|
|
|
by sgarland
615 days ago
|
|
|
> The updates wouldn’t be in place anyway because of the Postgres mvcc approach, right? They might be if the columns being updated weren’t indexed [0], but since it sounds like at least one is, no, not in-place. Though interestingly, your comment on BRIN indexes is quite relevant, as that’s the one type of index that HOT can still work with. [0]: https://www.postgresql.org/docs/current/storage-hot.html |
|
|