I am amused by the number written by Dr. Hipp. I appreciate his desire to keep core small and stable, but some of these would be great to get promoted into the base package.
News to me, there is a Python package[0], sqlean-py, which makes it trivial to bring in these extensions.
Oh no doubt, things are always improving. While I have not yet used it, happy for the recent json enhancements (personally, I would always prenormalize to standard JSON instead of naively ingesting JSON5).
Datetime is definitely one of my SQLite sore spots, so happy to see improvements there. I have to consult the docs + experimentation every time I need to do something with time manipulation. Even then, I usually left a bit disappointed with the available functionality.
Given that SQLite supports huge sizes for text, you could almost use the above as a way to create an outliner, with everything stored in the database itself ...
There is a valuable section about alternative tree models in the book "SQL Antipatterns" from Bill Karwin.
In the end, he compares five hierarchical data designs. Unsurprisingly, all have pros and cons and it is a (somehow advanced) matter of picking the right tool for the job. Closure tables may in fact be the lesser know variant, adjacency lists and recursive queries one of the better knows. But there are also path enumerations and nested sets.
> Bloom filter — a fast index to tell if a value is probably in a table or certainly isn't.
I will start using this way to communicate "true/false negative/positive" that always requires a further explanation.
Clear statements are part of the user experience.
News to me, there is a Python package[0], sqlean-py, which makes it trivial to bring in these extensions.
[0] https://antonz.org/sqlean-py/