Skip to content

Quality

expect_any_to_one(col1, col2)

A decorator function that ensures an N:1 relationship between col1 and col2, meaning each value in col1 corresponds to only one distinct value in col2.

Parameters:

Name Type Description Default
col1 str | Sequence[str]

Name of the column or a tuple of column names.

 required
col2 str | Sequence[str]

Name of the column or a tuple of column names.

 required
Source code in pysparky/quality.py 
 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
def expect_any_to_one(col1: str | Sequence[str], col2: str | Sequence[str]):
    """
    A decorator function that ensures an N:1 relationship between col1 and col2,
    meaning each value in col1 corresponds to only one distinct value in col2.

    Args:
        col1 (str | Sequence[str]): Name of the column or a tuple of column names.
        col2 (str | Sequence[str]): Name of the column or a tuple of column names.
    """

    def decorator(func):
        def wrapper(*args, **kwargs):
            spark_table_sdf = func(*args, **kwargs)
            num_col2_values_with_many_col1_values = (
                spark_table_sdf.groupBy(*col1)
                .agg(F.count_distinct(*col2).alias("distinct_count"))
                .where(F.col("distinct_count") > 1)
                .count()
            )
            assert (
                num_col2_values_with_many_col1_values == 0
            ), f"Multiple {col2}s per {col1}"
            print(f"✅: {col1}:{col2} is N:1")
            return spark_table_sdf

        return wrapper

    return decorator

expect_criteria(criteria)

A decorator function that ensures a specific criterion on a Spark DataFrame.

Parameters:

Name Type Description Default
criteria Column

The filter criterion to be applied to the DataFrame.

 required

Returns:

Name Type Description
function

A decorated function that checks the criterion.

Raises:

Type Description
AssertionError

If the filtered count and unfiltered count of the DataFrame are not equal.
Source code in pysparky/quality.py 
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
def expect_criteria(criteria):
    """
    A decorator function that ensures a specific criterion on a Spark DataFrame.

    Parameters:
        criteria (pyspark.sql.column.Column): The filter criterion to be applied to the DataFrame.

    Returns:
        function: A decorated function that checks the criterion.

    Raises:
        AssertionError: If the filtered count and unfiltered count of the DataFrame are not equal.
    """

    def decorator(func):
        def wrapper(*args, **kwargs):
            spark_table_sdf = func(*args, **kwargs)
            filtered_count = spark_table_sdf.filter(criteria).count()
            unfiltered_count = spark_table_sdf.count()
            assert (
                filtered_count == unfiltered_count
            ), f"Filtered count is not equal to unfiltered count {criteria}"
            print(f"✅: Criteria '{criteria}' passed")
            return spark_table_sdf

        return wrapper

    return decorator

expect_one_to_one(col1, col2)

A decorator function that ensures a 1:1 relationship between col1 and col2, meaning each value in col1 corresponds to only one distinct value in col2 and vice-versa.

Parameters:

Name Type Description Default
col1 str | Sequence[str]

Name of the column or a tuple of column names.

 required
col2 str | Sequence[str]

Name of the column or a tuple of column names.

 required
Source code in pysparky/quality.py 
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
def expect_one_to_one(col1: str | Sequence[str], col2: str | Sequence[str]):
    """
    A decorator function that ensures a 1:1 relationship between col1 and col2,
    meaning each value in col1 corresponds to only one distinct value in col2 and vice-versa.

    Args:
        col1 (str | Sequence[str]): Name of the column or a tuple of column names.
        col2 (str | Sequence[str]): Name of the column or a tuple of column names.
    """

    any_to_one = expect_any_to_one(col1, col2)
    one_to_any = expect_any_to_one(col2, col1)

    def decorator(func):
        return one_to_any(any_to_one(func))

    return decorator

expect_type(col_name, col_type)

A decorator function that verifies the data type of a specified column in a Spark DataFrame.

Parameters:

Name Type Description Default
col_name str

The column's name.

 required
col_type DataType

The expected data type for the column.

 required

Returns:

Name Type Description
function

A decorated function that checks the column's data type.

Raises:

Type Description
AssertionError

If the column's data type does not match the expected type.
Source code in pysparky/quality.py 
 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
def expect_type(col_name, col_type):
    """
    A decorator function that verifies the data type of a specified column in a Spark DataFrame.

    Parameters:
        col_name (str): The column's name.
        col_type (pyspark.sql.types.DataType): The expected data type for the column.

    Returns:
        function: A decorated function that checks the column's data type.

    Raises:
        AssertionError: If the column's data type does not match the expected type.
    """

    def decorator(func):
        def wrapper(*args, **kwargs):
            spark_table_sdf = func(*args, **kwargs)
            source_type = spark_table_sdf.schema[col_name].dataType
            target_type = col_type
            assert (
                source_type == target_type
            ), f"Data type of column '{col_name}:{source_type}' is not equal to {target_type}"  # noqa: E501
            print(f"✅: Column '{col_name}' has the expected data type {col_type}")
            return spark_table_sdf

        return wrapper

    return decorator

expect_unique(col_name)

A decorator function that ensures the uniqueness of a column in a Spark DataFrame.

Parameters:

Name Type Description Default
col_name str

The column's name.

 required

Returns:

Name Type Description
function

A decorated function that checks the column's uniqueness.

Raises:

Type Description
AssertionError

If the column's count and distinct count are not equal.
Source code in pysparky/quality.py 
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
def expect_unique(col_name):
    """
    A decorator function that ensures the uniqueness of a column in a Spark DataFrame.

    Parameters:
        col_name (str): The column's name.

    Returns:
        function: A decorated function that checks the column's uniqueness.

    Raises:
        AssertionError: If the column's count and distinct count are not equal.
    """

    def decorator(func):
        def wrapper(*args, **kwargs):
            spark_table_sdf = func(*args, **kwargs)
            spark_table_col_sdf = spark_table_sdf.select(col_name)
            normal_count = spark_table_col_sdf.count()
            distinct_count = spark_table_col_sdf.distinct().count()
            assert (
                normal_count == distinct_count
            ), f"Count and distinct count of column '{col_name}' are not equal"
            print(f"✅: Column '{col_name}' is distinct")
            return spark_table_sdf

        return wrapper

    return decorator