Skip to content
A python wrapper for the KSQL REST API.
Python Other
  1. Python 98.9%
  2. Other 1.1%
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Type Name Latest commit message Commit time
Failed to load latest commit information.
ksql-server update ksql version to 5.0.0-beta1 Jun 19, 2018
tests added a unit test for the query method Oct 2, 2018
.gitignore init Dec 9, 2017
.travis.yml remove old python version Aug 6, 2018
Dockerfile change to edge for latest librbkafka Oct 26, 2018
LICENSE init Dec 9, 2017
README.rst Added correct syntax highlighting for Python methods in the descriptions Jan 8, 2019
docker-compose.yaml implemented query with idle_timeout for better testing Sep 26, 2018
setup.cfg add cfg Dec 9, 2017 updated version Feb 24, 2019
test-requirements.txt merged changes from master Oct 2, 2018



A python wrapper for the KSQL REST API. Easily interact with the KSQL REST API using this library.

Supported KSQL version: 5.x


pip install ksql


git clone
cd ksql-python
python install

Getting Started

Setup for KSQL

This is the GITHUB page of KSQL.

If you have installed open source Confluent CLI (e.g. by installing Confluent Open Source or Enterprise Platform), you can start KSQL and its dependencies with one single command:

confluent start ksql-server

Setup for ksql-python API

  • Setup for the KSQL API:
from ksql import KSQLAPI
client = KSQLAPI('http://ksql-server:8088')
  • Setup for KSQl API with logging enabled:
import logging
from ksql import KSQLAPI
client = KSQLAPI('http://ksql-server:8088')


Option Type Required Description
url string yes Your ksql-server url. Example: http://ksql-server:8080
timeout integer no Timout for Requests. Default: 5

Main Methods


This method can be used for some KSQL features which are not supported via other specific methods like query, create_stream or create_stream_as. The following example shows how to execute the show tables statement:

client.ksql('show tables')
  • Example Response [{'tables': {'statementText': 'show tables;', 'tables': []}}]


It will execute sql query and keep listening streaming data.

client.query('select * from table1')

This command returns a generator. It can be printed e.g. by reading its values via next(query) or a for loop. Here is a complete example:

from ksql import KSQLAPI
client = KSQLAPI('http://localhost:8088')
query = client.query('select * from table1')
for item in query: print(item)
  • Example Response


Simplified API

create_stream/ create_table



Option Type Required Description
table_name string yes name of stream/table
columns_type list yes ex:['viewtime bigint','userid varchar','pageid varchar']
topic string yes Kafka topic
value_format string no JSON (Default) or DELIMITED or AVRO
key string for Table Key (used for JOINs)
  • Responses
If create table/stream succeed:return True
If failed:raise a CreateError(respose_from_ksql_server)


a simplified api for creating stream as select

CREATE STREAM <table_name>
[WITH ( kafka_topic=<kafka_topic>, value_format=<value_format>, property_name=expression ... )]
AS SELECT  <select_columns>
FROM <src_table>
[WHERE <conditions>]
PARTITION BY <partition_by>];


Option Type Required Description
table_name string yes name of stream/table
select_columns list yes you can select [*] or ['columnA', 'columnB']
src_table string yes name of source table
kafka_topic string no The name of the Kafka topic of this new stream(table).
value_format string no DELIMITED, JSON``(Default) or ``AVRO
conditions string no The conditions in the where clause.
partition_by string no Data will be distributed across partitions by this column.
kwargs pair no please provide key=value pairs. Please see more options.


KSQL JOINs between Streams and Tables are not supported yet via explicit methods, but you can use the ksql method for this like the following:

client.ksql("CREATE STREAM join_per_user WITH (VALUE_FORMAT='AVRO', KAFKA_TOPIC='join_per_user') AS SELECT Time, Amount FROM source c INNER JOIN users u on c.user = u.userid WHERE u.USERID = 1")



Run commands from a .ksql file. Can only support ksql commands and not streaming queries.

from ksql import FileUpload
pointer = FileUpload('http://ksql-server:8080')


Option Type Required Description
ksqlfile string yes name of file containing the rules
  • Responses
If ksql-commands succesfully executed:return (List of server response for all commands)
If failed:raise the appropriate error

More Options

There are more properties (partitions, replicas, etc...) in the official document.

KSQL Syntax Reference

  • Responses
If create table/stream succeed:return True
If failed:raise a CreatError(respose_from_ksql_server)
You can’t perform that action at this time.