cassarrow 0.1.0

An Apache Arrow adapter to Cassandra python driver

Cassarrow

Arrow based Cassandra python driver.

TLDR;

Speed up the cassandra python driver using C++ to parse cassandra queries data as Apache Arrow tables.

Key features:

• 20x speed up in the parsing of results
• 14x less memory
• Support for most native types, UDT, List and Set

Getting Started

Installation

pip install cassarrow

Usage

import cassarrow
import pyarrow as pa

# ...

with cassarrow.install_cassarrow(session) as cassarrow_session:
    table: pa.Table = cassarrow.result_set_to_table(cassarrow_session.execute_query("SELECT * FROM my_table"))

Type Mapping

Native Types

Cassandra	pyarrow	Note
ascii	pa.string()
bigint	pa.int64()
blob	pa.binary()
boolean	pa.bool_()
counter		Not supported
date	pa.date32()
decimal		Not supported
double	pa.float64()
duration	pa.duration("ns")
float	pa.float32()
inet		Not supported
int	pa.int32()
smallint	pa.int16()
text	pa.string()
time	pa.time64("ns")
timestamp	pa.timestamp("ms")
timeuuid	pa.binary(16)
tinyint	pa.int8()
uuid	pa.binary(16)
varchar	pa.string()
varint		Not supported

Collections / UDT

Cassandra	pyarrow	Note
list	pa.list_
map	pa.map_
set	pa.list_
udt	pa.struct