Fix TimestampGen to generate value not too close to the minimum allowed timestamp [databricks]

integration_tests/src/main/python/csv_test.py

@@ -402,7 +402,6 @@ def test_read_valid_and_invalid_dates(std_input_path, filename, v1_enabled_list,
         "'T'HH:mm[:ss]",
         "'T'HH:mm"]
 
-@datagen_overrides(seed=0, reason='https://github.com/NVIDIA/spark-rapids/issues/9701')
 @pytest.mark.parametrize('ts_part', csv_supported_ts_parts)
 @pytest.mark.parametrize('date_format', csv_supported_date_formats)
 @pytest.mark.parametrize('v1_enabled_list', ["", "csv"])

integration_tests/src/main/python/data_gen.py

@@ -578,9 +578,9 @@ def __init__(self, start=None, end=None, nullable=True, tzinfo=timezone.utc):
             # Spark supports times starting at
             # "0001-01-01 00:00:00.000000"
             # but it has issues if you get really close to that because it tries to do things
-            # in a different format which causes roundoff, so we have to add a few days,
+            # in a different format which causes roundoff, so we have to add a few days, even a month,
             # just to be sure
-            start = datetime(1, 1, 3, tzinfo=tzinfo)
+            start = datetime(1, 2, 1, tzinfo=tzinfo)
         elif not isinstance(start, datetime):
             raise RuntimeError('Unsupported type passed in for start {}'.format(start))
 

integration_tests/src/main/python/parquet_test.py

@@ -311,7 +311,6 @@ def test_parquet_pred_push_round_trip(spark_tmp_path, parquet_gen, read_func, v1
             lambda spark: rf(spark).select(f.col('a') >= s0),
             conf=all_confs)
 
-@datagen_overrides(seed=0, reason='https://github.com/NVIDIA/spark-rapids/issues/9701')
 @pytest.mark.parametrize('parquet_gens', [parquet_nested_datetime_gen], ids=idfn)
 @pytest.mark.parametrize('ts_type', parquet_ts_write_options)
 @pytest.mark.parametrize('ts_rebase_write', [('CORRECTED', 'LEGACY'), ('LEGACY', 'CORRECTED')])

integration_tests/src/main/python/parquet_write_test.py

@@ -458,7 +458,6 @@ def generate_map_with_empty_validity(spark, path):
             lambda spark, path: spark.read.parquet(path),
             data_path)
 
-@datagen_overrides(seed=0, reason='https://github.com/NVIDIA/spark-rapids/issues/9701')
 @pytest.mark.parametrize('data_gen', parquet_nested_datetime_gen, ids=idfn)
 @pytest.mark.parametrize('ts_write', parquet_ts_write_options)
 @pytest.mark.parametrize('ts_rebase_write', ['EXCEPTION'])
@@ -475,7 +474,6 @@ def writeParquetCatchException(spark, data_gen, data_path):
         lambda spark: writeParquetCatchException(spark, data_gen, data_path),
         conf=all_confs)
 
-@datagen_overrides(seed=0, reason='https://github.com/NVIDIA/spark-rapids/issues/9701')
 @pytest.mark.parametrize('data_gen', parquet_nested_datetime_gen, ids=idfn)
 @pytest.mark.parametrize('ts_write', parquet_ts_write_options)
 @pytest.mark.parametrize('ts_rebase_write', [('CORRECTED', 'LEGACY'), ('LEGACY', 'CORRECTED')])