def test_is_dataframe(self):
df = pd.DataFrame({
'category': ['a', 'b', 'c'],
'gender': ['F', 'M', 'F'],
'age': [1, 5, 4]
})
self.assertTrue(filters.is_dataframe(df))
def test_is_dataframe(self):
df = pd.DataFrame({
'category': ['a', 'b', 'c'],
'gender': ['F', 'M', 'F'],
'age': [1, 5, 4]
})
self.assertTrue(filters.is_dataframe(df))
def test_is_not_dataframe(self):
junk_inputs = [
'foo',
42,
[1, 2, 3],
[[1, 2, 3], [1, 2, 3], [1, 2, 3], ],
{'a': 1}
]
for junk in junk_inputs:
self.assertFalse(filters.is_dataframe(junk))
def test_is_not_dataframe(self):
junk_inputs = [
'foo',
42,
[1, 2, 3],
[[1, 2, 3], [1, 2, 3], [1, 2, 3], ],
{'a': 1}
]
for junk in junk_inputs:
self.assertFalse(filters.is_dataframe(junk))
def write_to_db_agnostic(engine, table, dataframe, schema=None):
"""
Given an sqlalchemy engine or sqlite connection, writes a dataframe to a table
Args:
engine (sqlalchemy.engine.base.Engine, sqlite3.Connection): the database engine or connection object
table (str): destination table
dataframe (pandas.DataFrame): the data to write
schema (str): the optional database schema
"""
# Validate inputs
is_engine = isinstance(engine, sqlalchemy.engine.base.Engine)
if sqlite3_is_loaded:
is_sqlite_connection = isinstance(engine, sqlite3.Connection)
else:
is_sqlite_connection = False
if not is_engine and not is_sqlite_connection:
raise HealthcareAIError('sqlalchemy engine or sqlite connection required, a {} was given'.format(type(engine)))
if not is_dataframe(dataframe):
raise HealthcareAIError('Dataframe required, a {} was given'.format(type(dataframe)))
if not isinstance(table, str):
raise HealthcareAIError('Table name required, a {} was given'.format(type(table)))
# Verify that tables exist for databases
if is_engine and not healthcareai.common.database_validators.does_table_exist(engine, table, schema):
raise HealthcareAIError('Destination table ({}) does not exist. Please create it.'.format(table))
elif is_sqlite_connection:
healthcareai.common.database_validators.verify_sqlite_table_exists(engine, table)
try:
# Count before
before_count = pd.read_sql('select count(*) from {}'.format(table), engine).iloc[0][0]
# Insert into database
dataframe.to_sql(table, engine, if_exists='append', index=False)
# Count after
after_count = pd.read_sql('select count(*) from {}'.format(table), engine).iloc[0][0]
delta = after_count - before_count
print('\nSuccessfully inserted {} rows. Dataframe contained {} rows'.format(delta, len(dataframe)))
# TODO catch other errors here:
except (sqlalchemy.exc.SQLAlchemyError, sqlite3.Error, pd.io.sql.DatabaseError):
raise HealthcareAIError("""Failed to insert values into {}.\n
Please verify that the table [{}] exists.\n
Was your test insert successful earlier?\n
If so, what has changed with your database/table/entity since then?""".format(table, table))
def write_to_db_agnostic(engine, table, dataframe, schema=None):
"""
Given an sqlalchemy engine or sqlite connection, writes a dataframe to a table
Args:
engine (sqlalchemy.engine.base.Engine, sqlite3.Connection): the database engine or connection object
table (str): destination table
dataframe (pandas.DataFrame): the data to write
schema (str): the optional database schema
"""
# Validate inputs
is_engine = isinstance(engine, sqlalchemy.engine.base.Engine)
if sqlite3_is_loaded:
is_sqlite_connection = isinstance(engine, sqlite3.Connection)
else:
is_sqlite_connection = False
if not is_engine and not is_sqlite_connection:
raise HealthcareAIError(
'sqlalchemy engine or sqlite connection required, a {} was given'.
format(type(engine)))
if not is_dataframe(dataframe):
raise HealthcareAIError('Dataframe required, a {} was given'.format(
type(dataframe)))
if not isinstance(table, str):
raise HealthcareAIError('Table name required, a {} was given'.format(
type(table)))
# Verify that tables exist for databases
if is_engine and not healthcareai.common.database_validators.does_table_exist(
engine, table, schema):
raise HealthcareAIError(
'Destination table ({}) does not exist. Please create it.'.format(
table))
elif is_sqlite_connection:
healthcareai.common.database_validators.verify_sqlite_table_exists(
engine, table)
try:
# Count before
before_count = pd.read_sql('select count(*) from {}'.format(table),
engine).iloc[0][0]
# Insert into database
dataframe.to_sql(table, engine, if_exists='append', index=False)
# Count after
after_count = pd.read_sql('select count(*) from {}'.format(table),
engine).iloc[0][0]
delta = after_count - before_count
print('\nSuccessfully inserted {} rows. Dataframe contained {} rows'.
format(delta, len(dataframe)))
# TODO catch other errors here:
except (sqlalchemy.exc.SQLAlchemyError, sqlite3.Error,
pd.io.sql.DatabaseError):
raise HealthcareAIError("""Failed to insert values into {}.\n
Please verify that the table [{}] exists.\n
Was your test insert successful earlier?\n
If so, what has changed with your database/table/entity since then?""".
format(table, table))