mini-ldw

A logical data warehouse demo. The supported databases & engines in this demo include:

MySQL (compatible with mysql-server-5.7.36, mysql-jdbc-5.1.41)
PostgreSQL (compatible with postgresql-server-9.5.19, postgresql-jdbc-42.3.1)
Clickhouse (compatible with clickhouse-server-22.1.3.7, clickhouse-jdbc-0.3.2-patch4)
Doris (compatible with doris-server-0.15.0-rc4, mysql-jdbc-5.1.41)
Elasticsearch (compatible with elasticsearch-7.10, elasticsearch-rest-client-7.10.2)

This logical data warehouse compatible with MySQL protocol and can be accessd with MySQL CLI and MySQL JDBC.

The goal of this project is simply to demonstrate the capabilities of Calcite in data virtualization, the metadata persistence and privilege management are not implemented. After server restarted, the schemas, tables and views that previously created will be lost.

This project currently supports creating view with views and tables in different data sources, this is the core feature of data virtualization product.

You can implement a customized data source adapter using Calcite Schema and Table related APIs. e.g., a graph database adapter.

Usage

Compile

mvn clean package -DskipTests

Start server

cd target && java -cp lib/*:mini-ldw-demo.jar com.mininglamp.MiniLDW

Access using MySQL CLI

mysql -h 127.0.0.1 -uroot -P9031 -A

Due to hard-to-locate bug, the -A parameter must be specified to disable auto-rehash.

Features

Create Schema

Create MySql schema

create schema mysql_schema 'mysql:127.0.0.1:3306:test_db:root:root';

Create PostgreSQL schema

create schema pg_schema 'postgresql:127.0.0.1:5432:test_db:root:root';

Create Clickhouse schema

create schema ck_schema 'clickhouse:127.0.0.1:8123:test_db:root:root';

Create Doris schema

create schema doris_schema 'mysql:127.0.0.1:9030:test_db:root:root';

Create Elasticsearch schema

create schema es_schema 'elasticsearch:127.0.0.1:9200:test_index:root:root';

If the index is not specified, then all indexes will be mapped to tables in the elasticsearch schema.

Drop Schema

drop schema mysql_schema;

Query

use mysql_schema;
select * from tbl where id = 1;

select * mysql_schema.tbl where id = 1;

Of course, it also supports multi-data-source fusion：

select t1.* from mysql_schema.tbl t1 join clickhouse_schema.tbl t2 on t1.object_id = t2.id where t2.name in (’happy‘, 'Chinese', 'new', 'year');

Create Table

create table tbl (id int, name varchar(50));

The data of this new creating table will be cached in heap memory, and will be lost after the server restarted.

Create View

create view test_view as select t1.* from mysql_schema.tbl t1 join clickhouse_schema.tbl t2 on t1.object_id = t2.id;

The data of this new creating view consists of multiple tables or views belonging to different data sources.

Insert & Update & Delete

insert into tbl (id, name) values (1, 'paradise');

update tbl set name = 'paradise' where id = 1;

delete from tbl where id = 1;

The insert, update and delete operations on tables in the physical data source will be persisted.

Show Schemas

show schemas;

Show Tables

show tables;

Describe Table

desc tbl;

SharpRay/mini-ldw