def _create_random_table(self, table_name, min_col_count, max_col_count,
allowed_storage_formats):
'''Create and return a Table with a random number of cols.'''
col_count = randint(min_col_count, max_col_count)
storage_format = choice(allowed_storage_formats)
table = Table(table_name)
table.storage_format = storage_format
allowed_types = list(TYPES)
# Avro doesn't support timestamps yet.
if table.storage_format == 'AVRO':
allowed_types.remove(Timestamp)
# TODO: 'table.cols' returns a copy of all scalar cols, so 'table.cols.append()'
# doesn't actually modify the table's columns. 'table.cols' should be changed
# to allow access to the real columns.
cols = table.cols
for col_idx in xrange(col_count):
col_type = choice(allowed_types)
col_type = choice(
filter(lambda type_: issubclass(type_, col_type), EXACT_TYPES))
if issubclass(col_type,
VarChar) and not issubclass(col_type, String):
col_type = get_varchar_class(randint(1, VarChar.MAX))
elif issubclass(col_type,
Char) and not issubclass(col_type, String):
col_type = get_char_class(randint(1, Char.MAX))
elif issubclass(col_type, Decimal):
max_digits = randint(1, Decimal.MAX_DIGITS)
col_type = get_decimal_class(max_digits,
randint(1, max_digits))
col = Column(
table, '%s_col_%s' % (col_type.__name__.lower(), col_idx + 1),
col_type)
cols.append(col)
table.cols = cols
return table
def create_random_table(self,
table_name,
min_number_of_cols,
max_number_of_cols,
allowed_storage_formats):
'''Create and return a Table with a random number of cols.'''
col_count = randint(min_number_of_cols, max_number_of_cols)
storage_format = choice(allowed_storage_formats)
table = Table(table_name)
table.storage_format = storage_format
for col_idx in xrange(col_count):
col_type = choice(TYPES)
col_type = choice(filter(lambda type_: issubclass(type_, col_type), EXACT_TYPES))
if issubclass(col_type, VarChar) and not issubclass(col_type, String):
col_type = get_varchar_class(randint(1, VarChar.MAX))
elif issubclass(col_type, Char) and not issubclass(col_type, String):
col_type = get_char_class(randint(1, Char.MAX))
elif issubclass(col_type, Decimal):
max_digits = randint(1, Decimal.MAX_DIGITS)
col_type = get_decimal_class(max_digits, randint(1, max_digits))
col = Column(
table,
'%s_col_%s' % (col_type.__name__.lower(), col_idx + 1),
col_type)
table.cols.append(col)
return table
def create_random_table(self, table_name, min_number_of_cols,
max_number_of_cols, allowed_storage_formats):
'''Create and return a Table with a random number of cols.'''
col_count = randint(min_number_of_cols, max_number_of_cols)
storage_format = choice(allowed_storage_formats)
table = Table(table_name)
table.storage_format = storage_format
for col_idx in xrange(col_count):
col_type = choice(TYPES)
col_type = choice(
filter(lambda type_: issubclass(type_, col_type), EXACT_TYPES))
if issubclass(col_type,
VarChar) and not issubclass(col_type, String):
col_type = get_varchar_class(randint(1, VarChar.MAX))
elif issubclass(col_type,
Char) and not issubclass(col_type, String):
col_type = get_char_class(randint(1, Char.MAX))
elif issubclass(col_type, Decimal):
max_digits = randint(1, Decimal.MAX_DIGITS)
col_type = get_decimal_class(max_digits,
randint(1, max_digits))
col = Column(
table, '%s_col_%s' % (col_type.__name__.lower(), col_idx + 1),
col_type)
table.cols.append(col)
return table
def _create_random_table(self,
table_name,
min_col_count,
max_col_count,
allowed_storage_formats):
'''Create and return a Table with a random number of cols.'''
col_count = randint(min_col_count, max_col_count)
storage_format = choice(allowed_storage_formats)
table = Table(table_name)
table.storage_format = storage_format
allowed_types = list(TYPES)
# Avro doesn't support timestamps yet.
if table.storage_format == 'AVRO':
allowed_types.remove(Timestamp)
# TODO: 'table.cols' returns a copy of all scalar cols, so 'table.cols.append()'
# doesn't actually modify the table's columns. 'table.cols' should be changed
# to allow access to the real columns.
cols = table.cols
for col_idx in xrange(col_count):
col_type = choice(allowed_types)
col_type = choice(filter(lambda type_: issubclass(type_, col_type), EXACT_TYPES))
if issubclass(col_type, VarChar) and not issubclass(col_type, String):
col_type = get_varchar_class(randint(1, VarChar.MAX))
elif issubclass(col_type, Char) and not issubclass(col_type, String):
col_type = get_char_class(randint(1, Char.MAX))
elif issubclass(col_type, Decimal):
max_digits = randint(1, Decimal.MAX_DIGITS)
col_type = get_decimal_class(max_digits, randint(1, max_digits))
col = Column(
table,
'%s_col_%s' % (col_type.__name__.lower(), col_idx + 1),
col_type)
cols.append(col)
table.cols = cols
return table
def FakeTable(name, fake_columns, storage_format='TEXTFILE'):
"""
Return a Table consisting of one or more FakeColumns. Because Columns are added via
method, we support nesting here instead.
"""
table = Table(name)
if not fake_columns:
raise Exception('You must supply at least one FakeColumn argument')
for fake_column in fake_columns:
table.add_col(fake_column)
table.storage_format = storage_format
return table
def test_table_model(self, cursor, hive_cursor):
table = Table("some_test_table")
cursor.drop_table(table.name, if_exists=True)
table.storage_format = 'textfile'
table.add_col(Column(table, "bigint_col", BigInt))
table.add_col(Column(table, "string_col", String))
cursor.create_table(table)
try:
other = hive_cursor.describe_table(table.name)
assert other.name == table.name
assert other.cols == table.cols
finally:
cursor.drop_table(table.name)
def test_table_model(self, cursor, hive_cursor):
table = Table("some_test_table")
cursor.drop_table(table.name, if_exists=True)
table.storage_format = 'textfile'
table.add_col(Column(table, "bigint_col", BigInt))
table.add_col(Column(table, "string_col", String))
cursor.create_table(table)
try:
other = hive_cursor.describe_table(table.name)
assert other.name == table.name
assert other.cols == table.cols
finally:
cursor.drop_table(table.name)
