Correct. That's the point. Postgres doesn't delete columns. Even after full vacuum (which is not something people do all the time), the damn column is still there.
Discussion
The column is there with absolutely no data and only remaining there as a tombstone. To remove it would require significant effort that has minimal benefit
It doesn't matter. The column is still there. That's what I said.
The column is there with what? Absolutely nothing! That column is about as good as gone even if you forensically analyse the files.
Like I said. It doesn't matter. The point of the post is that the column is still there, soft deleted.
Also, it's funny that you thought I didn't know any of this.
"The column" means the column's data. nostr:nprofile1qqs99d9qw67th0wr5xh05de4s9k0wjvnkxudkgptq8yg83vtulad30gpzamhxue69uhkuenyvghxummnwa5x2un99e3k7mgpp4mhxue69uhkummn9ekx7mqpz3mhxue69uhhyetvv9ujuerpd46hxtnfdujm6acf wins.
If X means Y, nobody wins.
Well, what remains about the column:
- a tombstone
What does not remain about it:
- all of its data
- the name and data type
- any indexes on it
Also, your original post very clearly 8: referring to data not being deleted, with the last remark on your post, and claiming the data is soft deleted.
Don’t look at CoW filesystems or filesystems in general, or flash translation layers, please