mkrd

- Instead of only indexing the key that was not found by the indexer, do a full indexing of all keys since full file was read anyways

```python DDB.at("test").create({"a": 0}, force_overwrite=True) with DDB.at("test", key="a").session() as (session, a): a = 1 session.write() print(DDB.at("test").read()) ``` prints 0, but it would be more intuitive if it was 1. The reason...

Currently, if a file does not exist, a FileNotFoundError is raised. If the file exists, but a key that does not exit was specified, then a KeyError is raised Proposal:...

Evaluate usage of mmap

1

mmap could greatly improve the performance of .at(…, key=…).session() - https://stackoverflow.com/questions/5917047/delete-insert-data-in-mmaped-file - https://stackoverflow.com/questions/50466557/python-write-content-into-existing-file - https://stackoverflow.com/questions/64145647/why-is-mmap-in-python-so-slow

Currently, a partial read still has to load the entire file into memory if the key was not found by the indexer Proposal: - `utils.find_outermost_key_in_json_bytes` and `seek_index_through_value_bytes` should be changed...

Change copy() spec. Instead of allowing the from_root argument, we should just have a .cop() method without any arguments. This will copy at the current path. and return a handle...

Find better examples

1

First, I appreciate the work you've put into this. But, frankly, I don't know if this is something I could see myself using. The main reason being that it introduces...

It might be worth documenting some of the edge cases, like what to use if any of the keys are tuples or the string "*", or how to insert a...