robin-sparkless 4.0.0

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
28
29
30
31
32
33
34
35
36
37
38
39
40
41

"""
Tests for issue #275: create_map() with no arguments (PySpark parity).

PySpark F.create_map() and F.create_map([]) return a column of empty maps ({} per row).
Robin previously raised: TypeError: py_create_map() missing 1 required positional argument: 'cols'
"""

from __future__ import annotations

from tests.fixtures.spark_imports import get_spark_imports

_imports = get_spark_imports()
F = _imports.F

from tests.utils import _row_to_dict


def test_create_map_no_args(spark) -> None:
    """create_map() with no args returns column of empty maps."""
    data = [{"id": 1}]
    df = spark.createDataFrame(data, ["id"])

    df = df.withColumn("m", F.create_map())
    rows = [_row_to_dict(r) for r in df.collect()]
    assert len(rows) == 1
    assert rows[0]["m"] is not None
    assert isinstance(rows[0]["m"], dict), "empty map should be dict {} (fixes #380)"
    assert rows[0]["m"] == {}


def test_create_map_empty_list(spark) -> None:
    """create_map([]) also returns column of empty maps (PySpark accepts both)."""
    data = [{"id": 1}, {"id": 2}]
    df = spark.createDataFrame(data, ["id"])

    df = df.withColumn("m", F.create_map())  # no args
    rows = [_row_to_dict(r) for r in df.collect()]
    assert len(rows) == 2
    for r in rows:
        assert isinstance(r["m"], dict), "empty map should be dict {} (fixes #380)"
        assert r["m"] == {}