Oracle

Oracle Connection

import connectorx as cx
conn = 'oracle://username:password@server:port/database'        # connection token
query = 'SELECT * FROM table'                                   # query string
cx.read_sql(conn, query)                                        # read data from Oracle

Oracle-Pandas Type Mapping

Oracle Type	Pandas Type	Comment
Number(*,0)	int64, Int64(nullable)
Number(*,>0)	float64
Float	float64
BINARY_FLOAT	float64
BINARY_DOUBLE	float64
VARCHAR2	object
CHAR	object
NCHAR	object
NVarchar2	object
DATE	datetime64[ns]
TIMESTAMP	datetime64[ns]
TIMESTAMP WITH TIME ZONE	datetime64[ns]

Performance (db.r5.4xlarge RDS)

Modin and Turbodbc does not support read_sql on Oracle

• Time chart, lower is better.

• Memory consumption chart, lower is better.

In conclusion, ConnectorX uses 3x less memory and 3x less time compared with Pandas.