.insert_all() and .upsert_all() to support iterator of Python lists/tuples

[simonw]

The table.insert_all() and table.upsert_all() methods currently only work with iterators of dictionaries, where each dictionary represents a row and the column names are the keys.

This is inefficient over large number of rows.

How about if they could also accept iterators that look like this?

def rows():
    yield "id", "title", "content"
    for row in some_other_source():
        yield row.id, row.title, row.content

This looks similar to CSV. The first "row" is the column titles, subsequent rows are the row values themselves.

[simonw]

I had Claude Code for web do a prototype of this idea, including benchmarks: https://github.com/simonw/research/blob/main/sqlite-utils-iterator-support/README.md

Prompt for that is in the PR:

• Add sqlite-utils iterator support research research#31

It looks good! Not a huge code change, reasonable speedup:

[simonw]

Documentation:

• https://github.com/simonw/sqlite-utils/blob/ecabd9e0c5f7e356c056f44f016d192d61d8caeb/docs/python-api.rst#inserting-data-from-a-list-or-tuple-iterator