Improve appender speed (1.1)

[JAicewizard]

I wasn't fully done with the PR, here are the implementations for parallel row and chunk sources (I will add parallel chunk tmrw).

The chunk version is, as expected, marginally better (~10%). I think this can be further improved by having faster chunk access in one way or another, as the bottleneck is still just setting the values and everything in involved with that

[JAicewizard]

@taniabogatsch I think the last remaining question is, is the schema actually usefull? Should it be kept in the API or removed.

Besides that I think this is ready for a review. (I will remove the draft once the schema issue is resolved)

[taniabogatsch]

Hi @JAicewizard. Thanks for the PR. This time, I'll not merge anything before it is ready. ;) We've just moved to v2 on main. Let me know once you've resolved all merge conflicts, and then I'll have a look at reviewing it and also at the schema question.

[JAicewizard]

I have a very busy couple of weeks ahead, generally this semester is quite full with projects so little time for this ATM. (I've tried to bump my duckdb extension to 1.2 for the past couple of weeks as well, that has been eating all my duckdb-time for now)

[taniabogatsch]

No worries - I hope the projects go well, and we'll get back to this PR whenever your schedule frees up again. :)