def test_insert_to_non_existent_table(self):
dataset = SparkHiveDataSet(database="default_1",
table="table_not_yet_created",
write_mode="insert")
dataset.save(_generate_spark_df_one())
assert_df_equal(dataset.load().sort("name"),
_generate_spark_df_one().sort("name"))
def test_read_from_non_existent_table(self):
dataset = SparkHiveDataSet(database="default_1",
table="table_doesnt_exist",
write_mode="insert")
with pytest.raises(
DataSetError,
match="requested table not found: default_1.table_doesnt_exist",
):
dataset.load()
def test_overwrite_empty_table(self, spark_hive_session):
spark_hive_session.sql(
"create table default_1.test_overwrite_empty_table (name string, age integer)"
).take(1)
dataset = SparkHiveDataSet(
database="default_1",
table="test_overwrite_empty_table",
write_mode="overwrite",
)
dataset.save(_generate_spark_df_one())
assert_df_equal(dataset.load(), _generate_spark_df_one())
def test_invalid_pk_provided(self):
dataset = SparkHiveDataSet(
database="default_1",
table="table_1",
write_mode="upsert",
table_pk=["column_doesnt_exist"],
)
with pytest.raises(
DataSetError,
match=r"Columns \[column_doesnt_exist\] selected as primary key\(s\) not "
r"found in table default_1\.table_1",
):
dataset.save(_generate_spark_df_one())
def test_upsert_empty_table(self, spark_session):
spark_session.sql(
"create table default_1.test_upsert_empty_table (name string, age integer)"
).take(1)
dataset = SparkHiveDataSet(
database="default_1",
table="test_upsert_empty_table",
write_mode="upsert",
table_pk=["name"],
)
dataset.save(_generate_spark_df_one())
assert_df_equal(
dataset.load().sort("name"), _generate_spark_df_one().sort("name")
)
def test_invalid_schema_insert(self, spark_hive_session):
spark_hive_session.sql(
"create table default_1.test_invalid_schema_insert "
"(name string, additional_column_on_hive integer)").take(1)
dataset = SparkHiveDataSet(
database="default_1",
table="test_invalid_schema_insert",
write_mode="insert",
)
with pytest.raises(
DataSetError,
match=r"dataset does not match hive table schema\.\n"
r"Present on insert only: \[\('age', 'int'\)\]\n"
r"Present on schema only: \[\('additional_column_on_hive', 'int'\)\]",
):
dataset.save(_generate_spark_df_one())
def test_upsert_config_err(self):
# no pk provided should prompt config error
with pytest.raises(
DataSetError,
match="table_pk must be set to utilise upsert read mode"):
SparkHiveDataSet(database="default_1",
table="table_1",
write_mode="upsert")
def test_cant_pickle(self):
import pickle # pylint: disable=import-outside-toplevel
with pytest.raises(pickle.PicklingError):
pickle.dumps(
SparkHiveDataSet(database="default_1",
table="table_1",
write_mode="overwrite"))
def test_invalid_write_mode_provided(self):
with pytest.raises(
DataSetError,
match="Invalid write_mode provided: not_a_write_mode"):
SparkHiveDataSet(
database="default_1",
table="table_1",
write_mode="not_a_write_mode",
table_pk=["name"],
)
def test_invalid_write_mode_provided(self):
pattern = (r"Invalid `write_mode` provided: not_a_write_mode\. "
r"`write_mode` must be one of: insert, upsert, overwrite")
with pytest.raises(DataSetError, match=pattern):
SparkHiveDataSet(
database="default_1",
table="table_1",
write_mode="not_a_write_mode",
table_pk=["name"],
)
def test_invalid_pk_provided(self):
with pytest.raises(
DataSetError,
match=r"columns \[column_doesnt_exist\] selected as PK not "
r"found in table default_1\.table_1",
):
SparkHiveDataSet(
database="default_1",
table="table_1",
write_mode="upsert",
table_pk=["column_doesnt_exist"],
)
def test_read_existing_table(self):
dataset = SparkHiveDataSet(database="default_1",
table="table_1",
write_mode="overwrite")
assert_df_equal(_generate_spark_df_one(), dataset.load())
