For 1), it is simple, but not trivial to parallelize csv parsing correctly, hence the yesterday’s post.

Otherwise, agree on 2 and 3. Object wrapping, number decoding can be slow on the whole “file to dataframe object” path, but trivially parallelizable.