Marcel Weisgut

Integrate Partial Indexes

14

This PR implements partial indexes, i.e., single-column indexes that can be created for one or more segments of the corresponding column but without the need to index all segments, and...

Update Third-Party Libraries

5

This PR updates used third-party libraries. **lib**|**version** ---|--- benchmark. | v1.7.0 compact_vector | ~~v0.1.1~~ 'compare' file not found -> use v0.1.0 ~until the CI uses ubuntu 22.04~ for now cxxopts...

Clean Up Wiki

1

In a team discussion, we decided to clean up our wiki. We only want to keep information that is relevant for getting started with Hyrise and that explains the tooling....

Investigate data races during concurrent chunk inserts

3

While working on #2453, I added a stress test, which concurrently appends chunks to a table. Executing the test with the thread sanitizer enabled results in data races. The data...

In a current CI run, the user is `root`, but the files and directories in the cloned Hyrise project directory have the owner `113`. Thus, we add the directory of...

Running `install_dependencies.sh` on a fresh macOS system results in the below warning: ``` DEPRECATION: Configuring installation scheme with distutils config files is deprecated and will no longer work in the...

Running `./scripts/coverage.sh` on macOS BigSur fails with llvm 12.0.1 ``` dyld: Symbol not found: ___llvm_profile_runtime Referenced from: [...]/hyrise/build-coverage/lib/libhyriseMvccDeletePlugin.dylib Expected in: flat namespace in [...]/hyrise/build-coverage/lib/libhyriseMvccDeletePlugin.dylib ./scripts/coverage.sh: line 51: 3419 Abort trap:...

Implement partial indexes. A first version is provided with #2386. Thanks to @bengelhaupt @Jasperhino @vxrahn. A copy of the code can be found at https://github.com/hyrise/hyrise/tree/dyod21/partial-hash-index. Further information about the motivation...

TPC-H Benchmark: item run limitation does not work in suffled mode

8

Executing the TPC-H benchmark in shuffled mode (`-m Shuffled`) with a runs-per-item limitation of `x` and a time limit that is high enough so that the query can be executed...

According to the contribution guidelines, `BaseVectorDecompressor` should actually be named `AbstractVectorDecompressor` since it is a base class for multiple differing implementations. It's the same with `BaseVectorCompressor`.