This is a basic and restricted implementation of jdbc driver for ClickHouse. It has support of a minimal subset of features to be usable.
<dependency>
<groupId>ru.yandex.clickhouse</groupId>
<artifactId>clickhouse-jdbc</artifactId>
<version>0.2</version>
</dependency>
URL syntax:
jdbc:clickhouse://<host>:<port>[/<database>]
, e.g. jdbc:clickhouse://localhost:8123/test
JDBC Driver Class:
ru.yandex.clickhouse.ClickHouseDriver
additionally, if you have a few instances, you can use BalancedClickhouseDataSource
.
In order to provide non-JDBC complaint data manipulation functionality, proprietary API exists.
Entry point for API is ClickHouseStatement#write()
method.
import ru.yandex.clickhouse.ClickHouseStatement;
ClickHouseStatement sth = connection.createStatement();
sth
.write() // Write API entrypoint
.table("default.my_table") // where to write data
.option("format_csv_delimiter", ";") // specific param
.data(new File("/path/to/file.csv"), ClickHouseFormat.CSV) // specify input
.send();
import ru.yandex.clickhouse.ClickHouseStatement;
ClickHouseStatement sth = connection.createStatement();
sth
.write()
.sql("INSERT INTO default.my_table (a,b,c)")
.data(new MyCustomInputStream(), ClickHouseFormat.JSONEachRow)
.addDbParam(ClickHouseQueryParam.MAX_PARALLEL_REPLICAS, 2)
.send();
import ru.yandex.clickhouse.ClickHouseStatement;
ClickHouseStatement sth = connection.createStatement();
sth.write().send("INSERT INTO test.writer", new ClickHouseStreamCallback() {
@Override
public void writeTo(ClickHouseRowBinaryStream stream) throws IOException {
for (int i = 0; i < 10; i++) {
stream.writeInt32(i);
stream.writeString("Name " + i);
}
}
},
ClickHouseFormat.RowBinary); // RowBinary or Native are supported
The driver is built with maven.
mvn package -DskipTests=true
To build a jar with dependencies use
mvn package assembly:single -DskipTests=true
In order to build the jdbc client one need to have jdk 1.6 or higher.