Speed up some functions

[ludwiglierhammer]

• cdm_mapper
• common.select

[ludwiglierhammer]

We used test data of 2.1 GB.
Reading (mdf_reader), selecting (common.select), mapping (cdm_mapper) and writing on disk were applied to the data:

Stage	chunksize	duration	max memory usage
Before	None	~120min	> 30GB
Before	200000	~40min	> 30GB
After	None	~50min	> 30GB
After	200000	~35min	> 30GB

This PR speeds up the code in general.
Unfortunately, maximum memory usage is not affected.

Memory rises (with chunking):

• at the end of mapping
• during selection (times 2 because of "data" and "mask")

[github-actions]

Warning
This Pull Request is coming from a fork and must be manually tagged approved
in order to perform additional testing.

[ludwiglierhammer]

@JanWillruth: For my perspective this PR is ready to merge. We can focus on the performance in another PR after we merged #348.