[ChrisRackauckas]

Repost from the Reddit:

https://h2oai.github.io/db-benchmark/ is obligatory to show as well. I think you should probably note that this isn't just DataFrames.jl but also in conjunction with Queryverse tools. the reason is because it's somewhat known that the Queryverse tools are very nice to use but not as performant as other parts of the Julia programming language (performance is a result of the language and how it's used). For example, the Parquet to DataFrame conversion that you're using has known performance issues: https://github.com/queryverse/ParquetFiles.jl/issues/32

In general, very nice benchmark contribution and thanks for helping showcase the performance landscape.