def test_column_fallback(sa):
engine = sa.create_engine("sqlite://")
data = pd.DataFrame({
"name": ["Frank", "Steve", "Jane", "Frank", "Michael"],
"age": [16, 21, 38, 22, 10],
"pet": ["fish", "python", "cat", "python", "frog"],
})
data.to_sql(name="test_sql_data", con=engine, index=False)
dataset = SqlAlchemyDataset("test_sql_data", engine=engine)
assert set(dataset.get_table_columns()) == {"name", "age", "pet"}
fallback_dataset = SqlAlchemyDataset("test_sql_data", engine=engine)
# override columns attribute to test fallback
fallback_dataset.columns = fallback_dataset.column_reflection_fallback()
assert set(fallback_dataset.get_table_columns()) == {"name", "age", "pet"}
# check that the results are the same for a few expectations
assert dataset.expect_column_to_exist(
"age") == fallback_dataset.expect_column_to_exist("age")
assert dataset.expect_column_mean_to_be_between(
"age",
min_value=10) == fallback_dataset.expect_column_mean_to_be_between(
"age", min_value=10)
# Test a failing expectation
assert dataset.expect_table_row_count_to_equal(
value=3) == fallback_dataset.expect_table_row_count_to_equal(value=3)
def test_column(sa):
engine = sa.create_engine("sqlite://")
data = pd.DataFrame(
{
"name": ["Frank", "Steve", "Jane", "Frank", "Michael"],
"age": [16, 21, 38, 22, 10],
"pet": ["fish", "python", "cat", "python", "frog"],
}
)
data.to_sql(name="test_sql_data", con=engine, index=False)
dataset = SqlAlchemyDataset("test_sql_data", engine=engine)
assert set(dataset.get_table_columns()) == {"name", "age", "pet"}
obs = dataset.columns
# Hacks to check instances of types
for col in obs:
col["type"] = str(col["type"])
assert len(obs) == 3
for column in [
{
"name": "name",
"type": "TEXT",
"nullable": True,
"default": None,
"autoincrement": "auto",
"primary_key": 0,
},
{
"name": "age",
"type": "BIGINT",
"nullable": True,
"default": None,
"autoincrement": "auto",
"primary_key": 0,
},
{
"name": "pet",
"type": "TEXT",
"nullable": True,
"default": None,
"autoincrement": "auto",
"primary_key": 0,
},
]:
assert column in obs
