def test_download_all_data_types(self):
# Test that all available data types from BigQuery (as of now)
# are handled properly
if not os.path.exists(self.bq_token):
raise nose.SkipTest(
'Skipped because authentication information is not available.')
query = """SELECT "PI" as VALID_STRING,
"" as EMPTY_STRING,
STRING(NULL) as NULL_STRING,
INTEGER(3) as VALID_INTEGER,
INTEGER(NULL) as NULL_INTEGER,
PI() as VALID_FLOAT,
FLOAT(NULL) as NULL_FLOAT,
TIMESTAMP("1970-01-01 00:00:00") as UNIX_EPOCH,
TIMESTAMP("2004-09-15 05:00:00") as VALID_TIMESTAMP,
TIMESTAMP(NULL) as NULL_TIMESTAMP,
BOOLEAN(TRUE) as TRUE_BOOLEAN,
BOOLEAN(FALSE) as FALSE_BOOLEAN,
BOOLEAN(NULL) as NULL_BOOLEAN"""
client = gbq._authenticate()
a = gbq.read_gbq(query,
col_order=[
'VALID_STRING', 'EMPTY_STRING', 'NULL_STRING',
'VALID_INTEGER', 'NULL_INTEGER', 'VALID_FLOAT',
'NULL_FLOAT', 'UNIX_EPOCH', 'VALID_TIMESTAMP',
'NULL_TIMESTAMP', 'TRUE_BOOLEAN', 'FALSE_BOOLEAN',
'NULL_BOOLEAN'
])
tm.assert_frame_equal(a, self.correct_test_datatype)
def test_table__not_exists(self):
# Test the inverse of `test_table_exists`
if not os.path.exists(self.bq_token):
raise nose.SkipTest('Skipped because authentication information is not available.')
client = gbq._authenticate()
table_reference = client.GetTableReference("publicdata:samples.does_not_exist")
self.assertFalse(client.TableExists(table_reference))
def test_download_dataset_larger_than_100k_rows(self):
# Test for known BigQuery bug in datasets larger than 100k rows
# http://stackoverflow.com/questions/19145587/bq-py-not-paging-results
if not os.path.exists(self.bq_token):
raise nose.SkipTest('Skipped because authentication information is not available.')
client = gbq._authenticate()
a = gbq.read_gbq("SELECT id, FROM [publicdata:samples.wikipedia] LIMIT 100005")
self.assertTrue(len(a) == 100005)
def test_table__not_exists(self):
# Test the inverse of `test_table_exists`
if not os.path.exists(self.bq_token):
raise nose.SkipTest(
'Skipped because authentication information is not available.')
client = gbq._authenticate()
table_reference = client.GetTableReference(
"publicdata:samples.does_not_exist")
self.assertFalse(client.TableExists(table_reference))
def test_download_dataset_larger_than_100k_rows(self):
# Test for known BigQuery bug in datasets larger than 100k rows
# http://stackoverflow.com/questions/19145587/bq-py-not-paging-results
if not os.path.exists(self.bq_token):
raise nose.SkipTest(
'Skipped because authentication information is not available.')
client = gbq._authenticate()
a = gbq.read_gbq(
"SELECT id, FROM [publicdata:samples.wikipedia] LIMIT 100005")
self.assertTrue(len(a) == 100005)
def test_table_exists(self):
# Given a table name in the format {dataset}.{tablename}, if a table exists,
# the GetTableReference should accurately indicate this.
# This could possibly change in future implementations of bq,
# but it is the simplest way to provide users with appropriate
# error messages regarding schemas.
if not os.path.exists(self.bq_token):
raise nose.SkipTest('Skipped because authentication information is not available.')
client = gbq._authenticate()
table_reference = client.GetTableReference("publicdata:samples.shakespeare")
self.assertTrue(client.TableExists(table_reference))
def test_table_exists(self):
# Given a table name in the format {dataset}.{tablename}, if a table exists,
# the GetTableReference should accurately indicate this.
# This could possibly change in future implementations of bq,
# but it is the simplest way to provide users with appropriate
# error messages regarding schemas.
if not os.path.exists(self.bq_token):
raise nose.SkipTest(
'Skipped because authentication information is not available.')
client = gbq._authenticate()
table_reference = client.GetTableReference(
"publicdata:samples.shakespeare")
self.assertTrue(client.TableExists(table_reference))
def test_unicode_string_conversion(self):
# Strings from BigQuery Should be converted to UTF-8 properly
if not os.path.exists(self.bq_token):
raise nose.SkipTest(
'Skipped because authentication information is not available.')
correct_test_datatype = DataFrame({'UNICODE_STRING': [u("\xe9\xfc")]})
query = """SELECT '\xc3\xa9\xc3\xbc' as UNICODE_STRING"""
client = gbq._authenticate()
a = gbq.read_gbq(query)
tm.assert_frame_equal(a, correct_test_datatype)
def test_unicode_string_conversion(self):
# Strings from BigQuery Should be converted to UTF-8 properly
if not os.path.exists(self.bq_token):
raise nose.SkipTest('Skipped because authentication information is not available.')
correct_test_datatype = DataFrame(
{'UNICODE_STRING' : [u("\xe9\xfc")]}
)
query = """SELECT '\xc3\xa9\xc3\xbc' as UNICODE_STRING"""
client = gbq._authenticate()
a = gbq.read_gbq(query)
tm.assert_frame_equal(a, correct_test_datatype)
def test_download_all_data_types(self):
# Test that all available data types from BigQuery (as of now)
# are handled properly
if not os.path.exists(self.bq_token):
raise nose.SkipTest('Skipped because authentication information is not available.')
query = """SELECT "PI" as VALID_STRING,
"" as EMPTY_STRING,
STRING(NULL) as NULL_STRING,
INTEGER(3) as VALID_INTEGER,
INTEGER(NULL) as NULL_INTEGER,
PI() as VALID_FLOAT,
FLOAT(NULL) as NULL_FLOAT,
TIMESTAMP("1970-01-01 00:00:00") as UNIX_EPOCH,
TIMESTAMP("2004-09-15 05:00:00") as VALID_TIMESTAMP,
TIMESTAMP(NULL) as NULL_TIMESTAMP,
BOOLEAN(TRUE) as TRUE_BOOLEAN,
BOOLEAN(FALSE) as FALSE_BOOLEAN,
BOOLEAN(NULL) as NULL_BOOLEAN"""
client = gbq._authenticate()
a = gbq.read_gbq(query, col_order = ['VALID_STRING',
'EMPTY_STRING',
'NULL_STRING',
'VALID_INTEGER',
'NULL_INTEGER',
'VALID_FLOAT',
'NULL_FLOAT',
'UNIX_EPOCH',
'VALID_TIMESTAMP',
'NULL_TIMESTAMP',
'TRUE_BOOLEAN',
'FALSE_BOOLEAN',
'NULL_BOOLEAN'])
tm.assert_frame_equal(a, self.correct_test_datatype)
