特征构造：OverWindow (OverWindowBatchOp)

Java 类名：com.alibaba.alink.operator.batch.feature.OverWindowBatchOp

Python 类名：OverWindowBatchOp

功能介绍

批式OverWindow特征构造组件。

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
clause	运算语句	运算语句	String	✓
orderBy	排序列	排序列	String	✓
groupCols	分组列名数组	分组列名，多列，可选，默认不选	String[]			null
reservedCols	算法保留列名	算法保留列	String[]			null

代码示例

Python 代码

from pyalink.alink import *

import pandas as pd

useLocalEnv(1)

df = pd.DataFrame([
    [2, 8, 2.2],
    [2, 7, 3.3],
    [5, 6, 4.4],
    [5, 6, 5.5],
    [7, 5, 6.6],
    [1, 8, 1.1],
    [1, 9, 1.0],
    [7, 5, 7.7],
    [9, 5, 8.8],
    [9, 4, 9.8],
    [19, 4, 8.8]])
data = BatchOperator.fromDataframe(df, schemaStr="f0 bigint, f1 bigint, f2 double")
OverWindowBatchOp().setOrderBy("f0, f1 desc").setClause("count_preceding(*) as cc").linkFrom(data).print()

Java 代码

import org.apache.flink.types.Row;

import com.alibaba.alink.operator.batch.BatchOperator;
import com.alibaba.alink.operator.batch.feature.OverWindowBatchOp;
import com.alibaba.alink.operator.batch.source.MemSourceBatchOp;
import org.junit.Test;

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

public class OverWindowBatchOpTest {
	@Test
	public void testOverWindowBatchOp() throws Exception {
		List <Row> df = Arrays.asList(
			Row.of(2, 8, 2.2),
			Row.of(2, 7, 3.3),
			Row.of(5, 6, 4.4),
			Row.of(5, 6, 5.5),
			Row.of(7, 5, 6.6),
			Row.of(1, 8, 1.1),
			Row.of(1, 9, 1.0),
			Row.of(7, 5, 7.7),
			Row.of(9, 5, 8.8),
			Row.of(9, 4, 9.8)
		);
		BatchOperator <?> data = new MemSourceBatchOp(df, "f0 int, f1 int, f2 double");
		new OverWindowBatchOp().setOrderBy("f0, f1 desc").setClause("count_preceding(*) as cc").linkFrom(data).print();
	}
}

运行结果

f0	f1	f2	cc
1	9	1.0000	0
1	8	1.1000	1
2	8	2.2000	2
2	7	3.3000	3
5	6	4.4000	4
5	6	5.5000	5
7	5	6.6000	6
7	5	7.7000	7
9	5	8.8000	8
9	4	9.8000	9