org.apache.beam.sdk.extensions.sql

Class BeamSql

java.lang.Object

org.apache.beam.sdk.extensions.sql.BeamSql

@Experimental
public class BeamSql
extends java.lang.Object

BeamSql is the DSL interface of BeamSQL. It translates a SQL query as a PTransform, so developers can use standard SQL queries in a Beam pipeline.

Beam SQL DSL usage:

A typical pipeline with Beam SQL DSL is:

 PipelineOptions options = PipelineOptionsFactory.create();
 Pipeline p = Pipeline.create(options);

 //create table from TextIO;
 PCollection<Row> inputTableA = p.apply(TextIO.read().from("/my/input/patha")).apply(...);
 PCollection<Row> inputTableB = p.apply(TextIO.read().from("/my/input/pathb")).apply(...);

 //run a simple query, and register the output as a table in BeamSql;
 String sql1 = "select MY_FUNC(c1), c2 from PCOLLECTION";
 PCollection<Row> outputTableA = inputTableA.apply(
    BeamSql
        .query(sql1)
        .registerUdf("MY_FUNC", MY_FUNC.class, "FUNC");

 //run a JOIN with one table from TextIO, and one table from another query
 PCollection<Row> outputTableB =
     PCollectionTuple
     .of(new TupleTag<>("TABLE_O_A"), outputTableA)
     .and(new TupleTag<>("TABLE_B"), inputTableB)
         .apply(BeamSql.query("select * from TABLE_O_A JOIN TABLE_B where ..."));

 //output the final result with TextIO
 outputTableB.apply(...).apply(TextIO.write().to("/my/output/path"));

 p.run().waitUntilFinish();
 

Constructor Summary

Constructors

Constructor and Description
BeamSql()

Method Summary

Modifier and Type	Method and Description
static QueryTransform	query(java.lang.String sqlQuery) Returns a QueryTransform representing an equivalent execution plan.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

BeamSql

public BeamSql()

Method Detail

query

public static QueryTransform query(java.lang.String sqlQuery)

Returns a QueryTransform representing an equivalent execution plan.

The QueryTransform can be applied to a PCollection or PCollectionTuple representing all the input tables.

The PTransform outputs a PCollection of Row.

If the PTransform is applied to PCollection then it gets registered with name PCOLLECTION.

If the PTransform is applied to PCollectionTuple then TupleTag.getId() is used as the corresponding PCollections name.

• If the sql query only uses a subset of tables from the upstream PCollectionTuple, this is valid;
• If the sql query references a table not included in the upstream PCollectionTuple, an IllegalStateException is thrown during query validation;
• Always, tables from the upstream PCollectionTuple are only valid in the scope of the current query call.