SQL操作：GroupBy (GroupByBatchOp)

Java 类名：com.alibaba.alink.operator.batch.sql.GroupByBatchOp

Python 类名：GroupByBatchOp

功能介绍

对批式数据进行sql的GROUPBY操作。

参数说明

名称	中文名称	描述	类型	是否必须？	取值范围	默认值
groupByPredicate	groupby语句	groupby语句	String	✓
selectClause	select语句	select语句	String	✓

代码示例

Python 代码

from pyalink.alink import *

import pandas as pd

useLocalEnv(1)

df = pd.DataFrame([
    ['Ohio', 2000, 1.5],
    ['Ohio', 2001, 1.7],
    ['Ohio', 2002, 3.6],
    ['Nevada', 2001, 2.4],
    ['Nevada', 2002, 2.9],
    ['Nevada', 2003, 3.2]
])

batch_data = BatchOperator.fromDataframe(df, schemaStr='f1 string, f2 bigint, f3 double')

op = GroupByBatchOp().setGroupByPredicate("f1").setSelectClause("f1,avg(f2) as f2")
batch_data = batch_data.link(op)

batch_data.print()

Java 代码

import org.apache.flink.types.Row;

import com.alibaba.alink.operator.batch.BatchOperator;
import com.alibaba.alink.operator.batch.source.MemSourceBatchOp;
import com.alibaba.alink.operator.batch.sql.GroupByBatchOp;
import org.junit.Test;

import java.util.Arrays;
import java.util.List;

public class GroupByBatchOpTest {
	@Test
	public void testGroupByBatchOp() throws Exception {
		List <Row> df = Arrays.asList(
			Row.of("Ohio", 2000, 1.5),
			Row.of("Ohio", 2001, 1.7),
			Row.of("Ohio", 2002, 3.6),
			Row.of("Nevada", 2001, 2.4),
			Row.of("Nevada", 2002, 2.9),
			Row.of("Nevada", 2003, 3.2)
		);
		BatchOperator <?> batch_data = new MemSourceBatchOp(df, "f1 string, f2 int, f3 double");
		BatchOperator <?> op = new GroupByBatchOp().setGroupByPredicate("f1").setSelectClause("f1,avg(f2) as f2");
		batch_data = batch_data.link(op);
		batch_data.print();
	}
}

运行结果

f1	f2
Nevada	2002
Ohio	2001