This is not to say vacuum-like mechanisms are necessarily the best method but it was a common architectural idiom for database engines from the 1990s because it works well with sequential storage devices. Newer techniques tend to put resource recovery inline with the workload rather than outside of it but that has other disadvantages.
Indexing all rows in a database by transaction id would cause an extreme loss of throughput. That just creates a continuously mutated (and therefore locking) structure every thread constantly uses that is also paged to disk. And deletion of individual records is a problem for indexes generally.
For example, as I recall, Oracle actually writes new row data to the page the row lives in, overwriting the old data in place on commit. It also writes the modifying transaction ID to the row header. If a different transaction finds the row, it will see that the transaction ID doesn't match (it's newer than itself), which forces the database to go to the undo log to look for the older data.
I'm not privy to the technical details of how the undo log is implemented or how it avoids the vacuum problem. I suspect it's related to the fact that the undo log is separate from the tuple data, so undo bloat doesn't affect table performance to the same extent as with Postgres.
In this case, I'd want to think carefully about the data volume: many places may have hundreds of transactions per second but also have queries which run for much longer periods of time and the visible state has to be accurate for all of them. That sounds a lot of contention for that shared data structure and “efficiently find all rows that aren't visible to any transaction” sounds decidedly non-trivial for busy servers with lots of data, which are generally the only ones where this matters.
Vacuum actually works by looking for tuples where no running transactions can see them anymore. Postgres, in effect, maintains a minimum and maximum transaction ID that any tuple is visible for, and vacuum scans over all of those that have a max visible transaction ID (suggesting it's available to be reclaimed) and which is less than the minimum active transaction ID.
Assuming active transactions {T}_i ordered by id, then if you are committing T_i, any row whose [Tmin, Tmax] visibility interval is fully contained in (T_(i-1), T_(i+1)) (i.e. for which Tmin > T_(i-1) && T_(i+1) < Tmax) is now dead (taking -inf and a large value for the previous/next transaction ids if there are none).
I believe this sort of query can be efficiently handled in time O(k polylog n) with several data structures, like an interval tree such as a B-tree augmented with maximum values or several kinds of 2D search trees.
It's also possible to use a simple index and just reclaim those rows whose Tmax is lower than the oldest active transaction, although this means that a single never-closed transaction blocks all row reclamation forever, which seems a bad design for a production-quality database.
Indexes in postgres point to the page, not the tuple. The VM keeps two bits per page: "does this page have any 'dead' tuples?", and "does this page need to be touched for xid wraparound?"
Vacuuming doesn't touch a page unless the VM indicates it's necessary.
Except for BRIN.
Note that it's not full scans - only pages that have been modified since the last vacuum, or were in a state last vacuum that they couldn't be processed, are vacuumed again. So there essentially is a block-level index for this.
https://www.postgresql.org/docs/current/static/storage-vm.ht...
It has potential pitfalls too though. On a system with very high mutation rate, purging could fall behind. This mainly would happen if the history list overflows available buffer pool space and starts getting paged to disk (e.g. from one very long lived transection). I don't know if newer MySQL versions make purge running off disk more efficient, but 5.5 added a config for multiple purge threads.
Many things can seem this way from the armchair, but you can lend some benefit of the doubt. Postgres is worked on by very smart people and it's not the only system that works this way
Everything has tradeoffs.
Postgres already has the notion of creating and destroying transaction ids for each row version attached to that tuple (row version). That distributes your "structure" across the rows involved in it. Just because you delete a row version (whether by deleting that row, or by updating it — which in MVCC is synchronously deleting the old version and inserting a new one) doesn't mean that row version isn't still visible to other concurrent activity, specifically including read-only activity. You'd therefore have to write to this proposed structure on completing every read or write.
Isn't it possible to just use a data structure to index rows by transaction ids, and on each transaction commit efficiently find all rows that aren't visible to any transaction, and add them to a free list?
Seems kind of a bad design to rely on periodic full data scans.