robin-sparkless 4.3.0

PySpark-like DataFrame API in Rust on Polars; no JVM.
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27

"""Tests for #418: createDataFrame with full nested DDL schema. Uses shared spark fixture."""

from __future__ import annotations


def test_nested_ddl_struct_and_array(spark) -> None:
    """createDataFrame(..., schema='addr struct<city:string>, tags array<string>') parses and creates schema."""
    # DDL with nested types: comma inside struct<> and array<> must not split top-level fields.
    schema_ddl = "addr struct<city:string>, tags array<string>"
    df = spark.createDataFrame(
        [
            {"addr": {"city": "NYC"}, "tags": ["a", "b"]},
            {"addr": {"city": "LA"}, "tags": ["c"]},
        ],
        schema=schema_ddl,
    )
    names = (
        df.columns if isinstance(getattr(df, "columns", None), list) else df.columns()
    )
    assert "addr" in names
    assert "tags" in names
    rows = df.collect()
    assert len(rows) == 2
    # Row content: addr as struct, tags as list
    r0 = rows[0]
    assert r0["addr"] is not None
    assert r0["tags"] is not None