Clarify docs and maybe refine example for larger than mem with duckdb

[drizk1]

The example right now fits into memory, but interestingly, when reading files from the url there's a much longer latency than if I copy them them to a local database db

This link outlines how to query larger than RAM files

SET memory_limit = '4GB';

SET temp_directory = '/tmp/duckdb_swap';
SET max_temp_directory_size = '100GB';

If you put a table name from the db directly in db_table it would do this after having it set up as above, it would work correctly.

I clarified the docs, and added a link. we could make a larger_than_mem equivalent that would automatically do this versus just leaving it as with the Julia / duckdb code in the example

Edit, did further reading and such, the slow down was in reading from a URL vs a local file. reverted most changes, kept the link and added note about URL vs local