Bouncner

Oh, they actually differ (see comment in `compressed_vector_type.hpp`). But maybe we can come up with better names for both that can be better distinguished and transfer more information.

@mrks just recommended `VectorCompressionAlgorithm` and `VectorCompressionType`. Since they are exactly the same for SimdBp128 and just differ in the actual byte-width for FSBA, I am not sure if those names...

Next time we should coordinate large PRs that change thousands of lines of code with rather simple changes. I am a little bit "afraid" of merge problems.

> It looks like there are many more unnecessary `hyrise::`s in the codebase :/ We should not block this PR just because we found something "new". But this PR should...

Putting it here: there are scans in the original queries that yield tuples in the paper data set but not in our data set. It appears to be a problem...

Yes, I cannot tell you if that really has an effect on the semi reduction performance BUT it's possible. We have a few very early semi joins that (due to...

One thing that might help here is to count (atomic_int or alikes) how many Insert operations are currently still being active (uncommitted, not yet rollbacked) per chunk.

Range Filters optimize for the gap sizes (maximize) between the histogram bins. Back in the days, we did not have a clue how to determine distances between strings. Now with...

I started implementing it a while ago and haven't had the time to finish the branch. Since there will be some work on this part in the near future, I...

Sounds good to me, but we should not underestimate the handling of predicate chains and the various characteristics of our indexes. I don't have a problem with dropping the B-Tree...