架构期望

使用 Great Expectations 的数据质量入门

Davina Moossazadeh

Data Scientist

形状与架构期望

"期望"（Expectation）——对数据的可验证断言

列期望
形状与架构期望
- 架构（schema）——数据集结构的蓝图

¹ https://docs.greatexpectations.io/docs/reference/learn/terms/expectation

行数

row_count_expectation = gx.expectations.ExpectTableRowCountToEqual(
    value=118000
)

validation_results = batch.validate(expect=row_count_expectation)

print(validation_results.success)

False

print(validation_results.result)

{'observed_value': 118066}

行数范围

使用 ExpectTableRowCountToBeBetween 为行数设定范围：

row_count_expectation = gx.expectations.ExpectTableRowCountToBeBetween(

    min_value=117000, max_value=119000
)

validation_results = batch.validate(expect=row_count_expectation)

print(validation_results.success)

True

print(validation_results.result)

{'observed_value': 118066}

列数

col_count_expectation = gx.expectations.ExpectTableColumnCountToEqual(
    value=15
)
validation_results = batch.validate(expect=col_count_expectation)

print(validation_results.success)

False

print(validation_result.result)

{'observed_value': 18}

列数范围

使用 ExpectTableColumnCountToBeBetween 为列数设定范围：

col_count_expectation = gx.expectations.ExpectTableColumnCountToBeBetween(
    min_value=14, max_value=18
)
validation_results = batch.validate(expect=col_count_expectation)

print(validation_results.success)

True

print(validation_result.result)

{'observed_value': 18}

列名集合

使用 ExpectTableColumnsToMatchSet 将列名与给定集合比对：

expected_cols = ['clouds_all', 'snow_1h', 'rain_1h', 'wind_speed', 'humidity', 
'pressure', 'temp', 'GHI', 'Energy delta[Wh]', 'Time', 'Time']

col_names_expectation = gx.expectations.ExpectTableColumnsToMatchSet(
    column_set=expected_cols
)

print(col_names_expectation.success, col_names_expectation.result)

True 
{'observed_value': ['Time', 'Energy delta[Wh]', 'GHI', 'temp', 'pressure',
'humidity', 'wind_speed', 'rain_1h', 'snow_1h', 'clouds_all'}

单列名称

使用 ExpectColumnToExist 检查数据集中是否存在指定列：

col_name_expectation = gx.expectations.ExpectColumnToExist(column="not_a_column")
validation_result = batch.validate(expect=col_name_expectation)
print(validation_result.success)

False

col_name_expectation = gx.expectations.ExpectColumnToExist(column="GHI")
validation_result = batch.validate(expect=col_name_expectation)
print(validation_result.success)

True

速查表

形状期望：

ExpectTableRowCountToEqual(value: int)
ExpectTableRowCountToBeBetween(
    min_value: int, max_value: int
)
ExpectTableColumnCountToEqual(
    value: int
)
ExpectTableColumnCountToBeBetween(
    min_value: int, max_value: int
)

列名期望：

ExpectTableColumnsToMatchSet(
    column_set: set
)
ExpectColumnToExist(column: str)

Passons à la pratique !

使用 Great Expectations 的数据质量入门