[edgarrmondragon]

Adds support for iterating map arrays:

# mre_map.py

# /// script
# requires-python = ">=3.10"
# dependencies = [
#     "nanoarrow",
#     "pyarrow",
# ]
# ///
"""Iterating a map array with nanoarrow fails."""

import nanoarrow as na
import pyarrow as pa

array = pa.array(
    [[("a", 1), ("b", None)], [], None, [("c", 3)]],
    type=pa.map_(pa.string(), pa.int64()),
)

print(f"nanoarrow {na.__version__}, pyarrow {pa.__version__}")
print(na.Array(array).to_pylist())

Run it with uv run --no-project mre_map.py. The output is:

nanoarrow 0.9.0, pyarrow 26.0.0
...
KeyError: "Can't resolve iterator for type 'map'"

[paleolimbot]

Thank you!

Just one correctness issue I think. I appreciate the contribution!

[paleolimbot]

I believe here we have to handle the zero length case first (before indexing into the entries)

[edgarrmondragon]

Thank you for the review :D. I added a quick return for the zero-length case at the top of the function.

[paleolimbot]

Thank you!