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
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
"""Tests for issue #259: datetime/date/Timestamp vs string comparisons.
Uses get_spark_imports from fixture only.
"""
from datetime import date, datetime
from typing import Iterable, List, Tuple
import pytest
from tests.fixtures.spark_imports import get_spark_imports
_imports = get_spark_imports()
SparkSession = _imports.SparkSession
F = _imports.F
class TestIssue259DatetimeStringComparison:
"""Regression tests for datetime/date/Timestamp vs string comparisons."""
def test_date_column_vs_string_column_filter(self) -> None:
"""Exact repro from GitHub issue #259 using date vs string columns."""
spark = SparkSession.builder.appName("Example").getOrCreate()
try:
df = spark.createDataFrame(
[
{
"Name": "Alice",
"date_timestamp": date(2026, 1, 1),
"date_string": "2025-06-15",
},
{
"Name": "Bob",
"date_timestamp": date(2026, 1, 1),
"date_string": "2026-01-30",
},
{
"Name": "Charlie",
"date_timestamp": date(2026, 1, 1),
"date_string": "2026-01-01",
},
]
)
# Compare date column to string column; should not raise TypeError
result = df.filter(F.col("date_timestamp") > F.col("date_string")).collect()
# PySpark expected:
# +-----+-----------+--------------+
# | Name|date_string|date_timestamp|
# +-----+-----------+--------------+
# |Alice| 2025-06-15| 2026-01-01|
# +-----+-----------+--------------+
assert len(result) == 1
assert result[0]["Name"] == "Alice"
assert result[0]["date_timestamp"] == date(2026, 1, 1)
assert result[0]["date_string"] == "2025-06-15"
finally:
spark.stop()
@pytest.mark.parametrize( # type: ignore[misc,untyped-decorator]
"op,expected_names",
[
(">", {"Alice"}),
(">=", {"Alice", "Charlie"}),
("<", {"Bob"}),
("<=", {"Bob", "Charlie"}),
("==", {"Charlie"}),
("!=", {"Alice", "Bob"}),
],
)
def test_date_column_vs_string_column_all_operators(
self, op: str, expected_names: Iterable[str]
) -> None:
"""Verify all comparison operators match PySpark semantics for date vs string columns."""
spark = SparkSession.builder.appName("Example").getOrCreate()
try:
df = spark.createDataFrame(
[
{
"Name": "Alice",
"d": date(2026, 1, 1),
"d_str": "2025-06-15",
},
{
"Name": "Bob",
"d": date(2026, 1, 1),
"d_str": "2026-01-30",
},
{
"Name": "Charlie",
"d": date(2026, 1, 1),
"d_str": "2026-01-01",
},
]
)
# Build comparison; handle each operator explicitly to match PySpark semantics.
if op == ">":
filtered = df.filter(F.col("d") > F.col("d_str"))
elif op == ">=":
filtered = df.filter(F.col("d") >= F.col("d_str"))
elif op == "<":
filtered = df.filter(F.col("d") < F.col("d_str"))
elif op == "<=":
filtered = df.filter(F.col("d") <= F.col("d_str"))
elif op == "==":
filtered = df.filter(F.col("d") == F.col("d_str"))
elif op == "!=":
filtered = df.filter(F.col("d") != F.col("d_str"))
else: # pragma: no cover - defensive
pytest.fail(f"Unsupported operator {op}")
names = {row["Name"] for row in filtered.collect()}
assert names == set(expected_names)
finally:
spark.stop()
def test_datetime_column_vs_string_column_filter(self) -> None:
"""Test datetime.datetime column compared to ISO datetime string column."""
spark = SparkSession.builder.appName("Example").getOrCreate()
try:
df = spark.createDataFrame(
[
{
"Name": "Early",
"ts": datetime(2025, 1, 1, 12, 0, 0),
"ts_str": "2024-12-31 23:59:59",
},
{
"Name": "Equal",
"ts": datetime(2025, 1, 1, 12, 0, 0),
"ts_str": "2025-01-01 12:00:00",
},
{
"Name": "Later",
"ts": datetime(2025, 1, 2, 0, 0, 0),
"ts_str": "2025-01-03 00:00:00",
},
]
)
# Keep rows where ts > ts_str
result = df.filter(F.col("ts") > F.col("ts_str")).collect()
names = {row["Name"] for row in result}
# Only "Early" should satisfy ts > ts_str
assert names == {"Early"}
finally:
spark.stop()
def test_datetime_column_vs_string_column_all_operators(self) -> None:
"""Verify all comparison operators for datetime vs string columns behave like PySpark."""
spark = SparkSession.builder.appName("Example").getOrCreate()
try:
base = datetime(2025, 1, 1, 12, 0, 0)
df = spark.createDataFrame(
[
{"Name": "Less", "ts": base, "ts_str": "2025-01-02 00:00:00"},
{"Name": "Equal", "ts": base, "ts_str": "2025-01-01 12:00:00"},
{"Name": "Greater", "ts": base, "ts_str": "2024-12-31 23:59:59"},
]
)
cases: List[Tuple[str, Iterable[str]]] = [
(">", {"Greater"}),
(">=", {"Greater", "Equal"}),
("<", {"Less"}),
("<=", {"Less", "Equal"}),
("==", {"Equal"}),
("!=", {"Less", "Greater"}),
]
for op, expected in cases:
if op == ">":
filtered = df.filter(F.col("ts") > F.col("ts_str"))
elif op == ">=":
filtered = df.filter(F.col("ts") >= F.col("ts_str"))
elif op == "<":
filtered = df.filter(F.col("ts") < F.col("ts_str"))
elif op == "<=":
filtered = df.filter(F.col("ts") <= F.col("ts_str"))
elif op == "==":
filtered = df.filter(F.col("ts") == F.col("ts_str"))
elif op == "!=":
filtered = df.filter(F.col("ts") != F.col("ts_str"))
else: # pragma: no cover - defensive
pytest.fail(f"Unsupported operator {op}")
names = {row["Name"] for row in filtered.collect()}
assert names == set(expected), f"Operator {op} mismatch"
finally:
spark.stop()
def test_date_column_vs_string_literal(self) -> None:
"""Test comparisons between date column and string literal thresholds."""
spark = SparkSession.builder.appName("Example").getOrCreate()
try:
df = spark.createDataFrame(
[
{"Name": "A", "d": date(2025, 6, 14)},
{"Name": "B", "d": date(2025, 6, 15)},
{"Name": "C", "d": date(2025, 6, 16)},
]
)
# Use string literal on the right-hand side
result_gt = df.filter(F.col("d") > "2025-06-15").collect()
names_gt = {row["Name"] for row in result_gt}
assert names_gt == {"C"}
# Use string literal on the left-hand side
result_lt = df.filter(F.col("d") < "2025-06-15").collect()
names_lt = {row["Name"] for row in result_lt}
assert names_lt == {"A"}
finally:
spark.stop()
def test_string_column_vs_date_literal_all_operators(self) -> None:
"""Ensure string column compared to date literal behaves like PySpark (coercion on string side)."""
spark = SparkSession.builder.appName("Example").getOrCreate()
try:
df = spark.createDataFrame(
[
{"Name": "A", "d_str": "2025-06-14"},
{"Name": "B", "d_str": "2025-06-15"},
{"Name": "C", "d_str": "2025-06-16"},
]
)
date_lit = date(2025, 6, 15)
cases: List[Tuple[str, Iterable[str]]] = [
(">", {"C"}),
(">=", {"B", "C"}),
("<", {"A"}),
("<=", {"A", "B"}),
("==", {"B"}),
("!=", {"A", "C"}),
]
for op, expected in cases:
if op == ">":
filtered = df.filter(F.col("d_str") > date_lit)
elif op == ">=":
filtered = df.filter(F.col("d_str") >= date_lit)
elif op == "<":
filtered = df.filter(F.col("d_str") < date_lit)
elif op == "<=":
filtered = df.filter(F.col("d_str") <= date_lit)
elif op == "==":
filtered = df.filter(F.col("d_str") == date_lit)
elif op == "!=":
filtered = df.filter(F.col("d_str") != date_lit)
else: # pragma: no cover - defensive
pytest.fail(f"Unsupported operator {op}")
names = {row["Name"] for row in filtered.collect()}
assert names == set(expected), f"Operator {op} mismatch"
finally:
spark.stop()
def test_invalid_or_null_date_strings_do_not_raise(self) -> None:
"""Invalid or null date strings should not raise and should behave like null comparisons."""
spark = SparkSession.builder.appName("Example").getOrCreate()
try:
df = spark.createDataFrame(
[
{
"Name": "Valid",
"d": date(2025, 6, 16),
"d_str": "2025-06-15",
},
{
"Name": "Invalid",
"d": date(2025, 6, 16),
"d_str": "not-a-date",
},
{
"Name": "NullString",
"d": date(2025, 6, 16),
"d_str": None,
},
]
)
# Filter where date > string; invalid/None strings should be treated as null comparisons
result = df.filter(F.col("d") > F.col("d_str")).collect()
names = {row["Name"] for row in result}
# Only the valid parseable date string should participate in comparison
assert names == {"Valid"}
finally:
spark.stop()
def test_string_values_that_look_like_dates_but_both_sides_string(self) -> None:
"""When both sides are strings, comparisons should remain lexicographic (no coercion)."""
spark = SparkSession.builder.appName("Example").getOrCreate()
try:
df = spark.createDataFrame(
[
{"Name": "A", "left": "2024-01-10", "right": "2024-01-2"},
{"Name": "B", "left": "2024-01-2", "right": "2024-01-10"},
]
)
# No date objects involved here, so string comparison semantics apply
result = df.filter(F.col("left") > F.col("right")).collect()
names = {row["Name"] for row in result}
# In lexicographic comparison, "2024-01-2" > "2024-01-10"
assert names == {"B"}
finally:
spark.stop()