def test_should_pass_empty_df_if_there_are_no_rules(spark_session):
df = empty_string_df(spark_session)
result = ValidateSparkDataFrame(spark_session, df).execute()
AssertValidationResult(column_name="col1", constraint_name="") \
.check(
actual=result,
expected_correct=df,
expected_erroneous=df
)
def test_should_pass_df_if_there_are_no_rules(spark_session):
df = spark_session.createDataFrame([["abc"], ["def"]],
schema=single_string_column_schema)
result = ValidateSparkDataFrame(spark_session, df).execute()
AssertValidationResult(column_name="col1", constraint_name="") \
.check(
actual=result,
expected_correct=df,
expected_erroneous=empty_string_df(spark_session)
)
def test_should_return_df_without_changes_if_all_rows_are_unique(spark_session):
df = spark_session.createDataFrame([["abc"], ["def"], ["ghi"]], schema=single_string_column_schema)
result = ValidateSparkDataFrame(spark_session, df) \
.is_unique("col1") \
.execute()
AssertValidationResult(column_name="col1", constraint_name="unique") \
.check(
actual=result,
expected_correct=df,
expected_erroneous=empty_string_df(spark_session)
)
def test_should_return_df_without_changes_if_empty_df_with_is_unique_constraint(spark_session):
df = empty_string_df(spark_session)
result = ValidateSparkDataFrame(spark_session, df) \
.is_unique("col1") \
.execute()
AssertValidationResult(column_name="col1", constraint_name="unique") \
.check(
actual=result,
expected_correct=df,
expected_erroneous=df
)
def test_should_pass_empty_df_with_not_null_constraint(spark_session):
df = empty_string_df(spark_session)
result = ValidateSparkDataFrame(spark_session, df) \
.is_not_null("col1") \
.execute()
AssertValidationResult(column_name="col1", constraint_name="not_null") \
.check(
actual=result,
expected_correct=df,
expected_erroneous=df
)
def test_should_return_df_without_changes_if_regex_matches_the_text(spark_session):
df = spark_session.createDataFrame([["abc"], ["def"], ["ghi"]], schema=single_string_column_schema)
result = ValidateSparkDataFrame(spark_session, df) \
.text_matches_regex("col1", ".*") \
.execute()
AssertValidationResult(column_name="col1", constraint_name="regex_match") \
.check(
actual=result,
expected_correct=df,
expected_erroneous=empty_string_df(spark_session)
)
def test_should_reject_all_rows_if_none_of_them_is_in_the_list(spark_session):
df = spark_session.createDataFrame([["abc"], ["a"], ["abcdefghi"]], schema=single_string_column_schema)
expected_errors = spark_session.createDataFrame([["abc"], ["a"], ["abcdefghi"]], schema=single_string_column_schema)
result = ValidateSparkDataFrame(spark_session, df) \
.one_of("col1", ["ab", "b"]) \
.execute()
AssertValidationResult(column_name="col1", constraint_name="one_of") \
.check(
actual=result,
expected_correct=empty_string_df(spark_session),
expected_erroneous=expected_errors
)
def test_should_reject_all_rows_if_regex_match_fails(spark_session):
df = spark_session.createDataFrame([["abc"], ["a"], ["abcdefghi"]], schema=single_string_column_schema)
expected_errors = spark_session.createDataFrame([["abc"], ["a"], ["abcdefghi"]], schema=single_string_column_schema)
result = ValidateSparkDataFrame(spark_session, df) \
.text_matches_regex("col1", "[0-9]+") \
.execute()
AssertValidationResult(column_name="col1", constraint_name="regex_match") \
.check(
actual=result,
expected_correct=empty_string_df(spark_session),
expected_erroneous=expected_errors
)
def test_should_return_df_without_changes_if_all_are_longer_than_lower_bound(
spark_session):
df = spark_session.createDataFrame([["abcdef"], ["ghijkl"]],
schema=single_string_column_schema)
result = ValidateSparkDataFrame(spark_session, df) \
.has_length_between("col1", 5, 20) \
.execute()
AssertValidationResult(column_name="col1", constraint_name="text_length") \
.check(
actual=result,
expected_correct=df,
expected_erroneous=empty_string_df(spark_session)
)
def test_should_reject_all_rows_if_all_are_too_short_or_too_long(
spark_session):
df = spark_session.createDataFrame([["abc"], ["a"], ["abcdefghi"]],
schema=single_string_column_schema)
expected_errors = spark_session.createDataFrame(
[["abc"], ["a"], ["abcdefghi"]], schema=single_string_column_schema)
result = ValidateSparkDataFrame(spark_session, df) \
.has_length_between("col1", 5, 8) \
.execute()
AssertValidationResult(column_name="col1", constraint_name="text_length") \
.check(
actual=result,
expected_correct=empty_string_df(spark_session),
expected_erroneous=expected_errors
)
def test_should_throw_error_if_constraint_is_not_a_numeric_column(
spark_session):
with pytest.raises(ValueError):
ValidateSparkDataFrame(spark_session, empty_string_df(spark_session)) \
.mean_column_value("col1", 10, 10) \
.execute()
def test_should_throw_error_if_lower_bound_is_greater_than_upper_bound(
spark_session):
with pytest.raises(ValueError):
ValidateSparkDataFrame(spark_session, empty_string_df(spark_session)) \
.has_length_between("col1", 10, 5) \
.execute()
def test_should_throw_error_if_there_are_duplicate_constraints(spark_session):
with pytest.raises(ValueError):
ValidateSparkDataFrame(spark_session, empty_string_df(spark_session)) \
.has_length_between("col1", 0, 10) \
.has_length_between("col1", 0, 5) \
.execute()
def test_should_throw_error_if_constraint_uses_non_existing_column(
spark_session):
with pytest.raises(ValueError):
ValidateSparkDataFrame(spark_session, empty_string_df(spark_session)) \
.has_length_between("column_that_does_not_exist", 0, 1) \
.execute()
def test_should_throw_error_if_there_are_duplicate_constraints(spark_session):
with pytest.raises(ValueError):
ValidateSparkDataFrame(spark_session, empty_string_df(spark_session)) \
.is_unique("col1") \
.is_unique("col1") \
.execute()
def test_should_throw_error_if_there_are_duplicate_constraints(spark_session):
with pytest.raises(ValueError):
ValidateSparkDataFrame(spark_session, empty_string_df(spark_session)) \
.text_matches_regex("column_that_does_not_exist", '.*') \
.text_matches_regex("column_that_does_not_exist", '[a-z]*') \
.execute()
