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
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238

"""
Tests for issue #437: F.mean() on string columns with numeric content.

PySpark interprets string columns as numeric for mean/avg; sparkless previously
raised TypeError: can't convert type 'str' to numerator/denominator when using
F.mean() on string columns. Sparkless now coerces string columns to numeric
(PySpark parity).
"""

import math

from tests.fixtures.spark_imports import get_spark_imports


def _norm(val):
    """Normalize for backend-agnostic assertion."""
    if val is None:
        return None
    if isinstance(val, (int, float)):
        return float(val) if isinstance(val, float) else int(val)
    return val


class TestIssue437MeanStringColumn:
    """Test F.mean() on string columns with numeric content (issue #437)."""

    def test_mean_string_column_exact_issue_scenario(self, spark):
        """Exact scenario from issue #437: F.mean() on string X and int Y."""
        imports = get_spark_imports()
        F = imports.F
        df = spark.createDataFrame(
            [
                {"Name": "Alice", "X": "1", "Y": 1},
                {"Name": "Alice", "X": "2", "Y": 2},
                {"Name": "Alice", "X": "3", "Y": 3},
                {"Name": "Bob", "X": "4", "Y": 4},
            ]
        )
        df = df.groupBy("Name").agg(
            F.mean(F.col("X")).alias("avg(X)"),
            F.mean(F.col("Y")).alias("avg(Y)"),
        )
        rows = df.collect()
        assert len(rows) == 2
        alice = next(r for r in rows if r["Name"] == "Alice")
        bob = next(r for r in rows if r["Name"] == "Bob")
        assert _norm(alice["avg(X)"]) == 2.0
        assert _norm(alice["avg(Y)"]) == 2.0
        assert _norm(bob["avg(X)"]) == 4.0
        assert _norm(bob["avg(Y)"]) == 4.0

    def test_mean_string_column_groupby(self, spark):
        """F.mean() on string column with groupBy."""
        imports = get_spark_imports()
        F = imports.F
        df = spark.createDataFrame(
            [
                {"Type": "A", "Value": "10"},
                {"Type": "A", "Value": "20"},
                {"Type": "A", "Value": "30"},
                {"Type": "B", "Value": "5"},
                {"Type": "B", "Value": "15"},
            ]
        )
        df = df.groupBy("Type").agg(F.mean("Value"))
        rows = df.collect()
        a_row = next(r for r in rows if r["Type"] == "A")
        b_row = next(r for r in rows if r["Type"] == "B")
        assert _norm(a_row["avg(Value)"]) == 20.0
        assert _norm(b_row["avg(Value)"]) == 10.0

    def test_mean_string_column_decimal_values(self, spark):
        """F.mean() on string column with decimal-like values."""
        imports = get_spark_imports()
        F = imports.F
        df = spark.createDataFrame(
            [
                {"Type": "A", "Value": "1.5"},
                {"Type": "A", "Value": "2.5"},
                {"Type": "A", "Value": "3.5"},
            ]
        )
        df = df.groupBy("Type").agg(F.mean("Value"))
        rows = df.collect()
        assert len(rows) == 1
        assert abs(_norm(rows[0]["avg(Value)"]) - 2.5) < 1e-9

    def test_mean_string_column_with_nulls(self, spark):
        """F.mean() on string column with nulls - nulls excluded from mean."""
        imports = get_spark_imports()
        F = imports.F
        df = spark.createDataFrame(
            [
                {"Type": "A", "Value": "10"},
                {"Type": "A", "Value": None},
                {"Type": "A", "Value": "20"},
                {"Type": "A", "Value": "30"},
            ]
        )
        df = df.groupBy("Type").agg(F.mean("Value"))
        rows = df.collect()
        assert len(rows) == 1
        # (10 + 20 + 30) / 3 = 20.0
        assert _norm(rows[0]["avg(Value)"]) == 20.0

    def test_mean_string_column_single_row_per_group(self, spark):
        """F.mean() on string column with single row per group."""
        imports = get_spark_imports()
        F = imports.F
        df = spark.createDataFrame(
            [
                {"Type": "A", "Value": "42"},
                {"Type": "B", "Value": "100"},
            ]
        )
        df = df.groupBy("Type").agg(F.mean("Value"))
        rows = df.collect()
        assert len(rows) == 2
        a_row = next(r for r in rows if r["Type"] == "A")
        b_row = next(r for r in rows if r["Type"] == "B")
        assert _norm(a_row["avg(Value)"]) == 42.0
        assert _norm(b_row["avg(Value)"]) == 100.0

    def test_avg_string_column_same_as_mean(self, spark):
        """F.avg() on string column - same behavior as F.mean()."""
        imports = get_spark_imports()
        F = imports.F
        df = spark.createDataFrame(
            [
                {"Type": "A", "Value": "1"},
                {"Type": "A", "Value": "2"},
                {"Type": "A", "Value": "3"},
            ]
        )
        df = df.groupBy("Type").agg(F.avg("Value"))
        rows = df.collect()
        assert len(rows) == 1
        assert _norm(rows[0]["avg(Value)"]) == 2.0

    def test_mean_string_column_multiple_aggregations(self, spark):
        """F.mean() + F.sum() + F.count() on string column."""
        imports = get_spark_imports()
        F = imports.F
        df = spark.createDataFrame(
            [
                {"Type": "A", "Value": "10"},
                {"Type": "A", "Value": "20"},
                {"Type": "A", "Value": "30"},
                {"Type": "B", "Value": "5"},
            ]
        )
        df = df.groupBy("Type").agg(
            F.mean("Value").alias("avg_val"),
            F.sum("Value").alias("sum_val"),
            F.count("Value").alias("cnt_val"),
        )
        rows = df.collect()
        a_row = next(r for r in rows if r["Type"] == "A")
        b_row = next(r for r in rows if r["Type"] == "B")
        assert _norm(a_row["avg_val"]) == 20.0
        assert _norm(a_row["sum_val"]) == 60.0
        assert a_row["cnt_val"] == 3
        assert _norm(b_row["avg_val"]) == 5.0
        assert _norm(b_row["sum_val"]) == 5.0
        assert b_row["cnt_val"] == 1

    def test_mean_string_column_select_after(self, spark):
        """F.mean() on string column, then select."""
        imports = get_spark_imports()
        F = imports.F
        df = spark.createDataFrame(
            [
                {"Type": "A", "Value": "10"},
                {"Type": "A", "Value": "20"},
                {"Type": "B", "Value": "5"},
            ]
        )
        result = df.groupBy("Type").agg(F.mean("Value")).select("Type", "avg(Value)")
        rows = result.collect()
        assert len(rows) == 2
        a_row = next(r for r in rows if r["Type"] == "A")
        b_row = next(r for r in rows if r["Type"] == "B")
        assert _norm(a_row["avg(Value)"]) == 15.0
        assert _norm(b_row["avg(Value)"]) == 5.0

    def test_mean_string_column_scientific_notation(self, spark):
        """F.mean() on string column with scientific notation (PySpark parity)."""
        imports = get_spark_imports()
        F = imports.F
        df = spark.createDataFrame(
            [
                {"Type": "A", "Value": "1e2"},
                {"Type": "A", "Value": "2e2"},
                {"Type": "A", "Value": "3e2"},
            ]
        )
        df = df.groupBy("Type").agg(F.mean("Value"))
        rows = df.collect()
        assert len(rows) == 1
        # 100 + 200 + 300 = 600, mean = 200
        assert abs(_norm(rows[0]["avg(Value)"]) - 200.0) < 1e-6

    def test_mean_string_column_mixed_int_float_strings(self, spark):
        """F.mean() on string column with mixed '1' and '1.0' style values."""
        imports = get_spark_imports()
        F = imports.F
        df = spark.createDataFrame(
            [
                {"Type": "A", "Value": "1"},
                {"Type": "A", "Value": "2.0"},
                {"Type": "A", "Value": "3"},
            ]
        )
        df = df.groupBy("Type").agg(F.mean("Value"))
        rows = df.collect()
        assert len(rows) == 1
        assert abs(_norm(rows[0]["avg(Value)"]) - 2.0) < 1e-9

    def test_mean_string_column_all_nulls_returns_none(self, spark):
        """F.mean() on string column with all nulls in group returns None/NaN."""
        imports = get_spark_imports()
        F = imports.F
        df = spark.createDataFrame(
            [
                {"Type": "A", "Value": "10"},
                {"Type": "B", "Value": None},
                {"Type": "B", "Value": None},
            ]
        )
        df = df.groupBy("Type").agg(F.mean("Value"))
        rows = df.collect()
        a_row = next(r for r in rows if r["Type"] == "A")
        b_row = next(r for r in rows if r["Type"] == "B")
        assert _norm(a_row["avg(Value)"]) == 10.0
        # PySpark returns null for empty aggregate; sparkless may return None
        assert b_row["avg(Value)"] is None or (
            isinstance(b_row["avg(Value)"], float) and math.isnan(b_row["avg(Value)"])
        )