Java 类名:com.alibaba.alink.operator.batch.dataproc.vector.VectorFunctionBatchOp
Python 类名:VectorFunctionBatchOp
功能介绍
- 获取一个向量的最大值、最小值,或者最大值、最小值的索引,或者对向量做尺度变换, 求NormL2, 求NormL1, 求NormL2Square, Normalize。
- 支持稀疏和稠密两种 Vector。
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 | | —- | —- | —- | —- | —- | —- | —- | | funcName | 函数名字 | 函数操作名称, 可取max(最大值), min(最小值), argMax(最大值索引), argMin(最小值索引), scale(尺度变换), NormL2, NormL1, NormL2Square, Normalize | String | ✓ | “Max”, “Min”, “ArgMax”, “ArgMin”, “Scale”, “NormL2”, “NormL1”, “NormL2Square”, “Normalize” | | | selectedCol | 选中的列名 | 计算列对应的列名 | String | ✓ | 所选列类型为 [DENSE_VECTOR, SPARSE_VECTOR, STRING, VECTOR] | | | WithVariable | Not available! | Not available! | String | | | | | outputCol | 输出结果列 | 输出结果列列名,可选,默认null | String | | | null | | reservedCols | 算法保留列名 | 算法保留列 | String[] | | | null | | numThreads | 组件多线程线程个数 | 组件多线程线程个数 | Integer | | | 1 |
代码示例
Python 代码
from pyalink.alink import *
import pandas as pd
useLocalEnv(1)
df = pd.DataFrame([
[1,"16.3, 1.1, 1.1"],
[2,"16.8, 1.4, 1.5"],
[3,"19.2, 1.7, 1.8"],
[4,"10.0, 1.7, 1.7"],
[5,"19.5, 1.8, 1.9"],
[6,"20.9, 1.8, 1.8"],
[7,"21.1, 1.9, 1.8"],
[8,"20.9, 2.0, 2.1"],
[9,"20.3, 2.3, 2.4"],
[10,"22.0, 2.4, 2.5"]
])
opData = BatchOperator.fromDataframe(df, schemaStr="id bigint, vec string")
result = VectorFunctionBatchOp().setSelectedCol("vec")\
.setOutputCol("out").setFuncName("max").linkFrom(opData)
result.collectToDataframe()
Java 代码
import org.apache.flink.types.Row;
import com.alibaba.alink.operator.stream.BatchOperator;
import com.alibaba.alink.operator.stream.dataproc.vector.VectorFunctionBatchOp;
import com.alibaba.alink.operator.stream.source.MemSourceBatchOp;
import com.alibaba.alink.testutil.AlinkTestBase;
import org.junit.Test;
import java.util.ArrayList;
import java.util.List;
public class VectorFunctionTest extends AlinkTestBase {
@Test
public void testVectorFunction() throws Exception {
List <Row> df = new ArrayList <>();
df.add(Row.of(1, "16.3, 1.1, 1.1"));
df.add(Row.of(2, "16.8, 1.4, 1.5"));
df.add(Row.of(3, "19.2, 1.7, 1.8"));
df.add(Row.of(4, "10.0, 1.7, 1.7"));
df.add(Row.of(5, "19.5, 1.8, 1.9"));
df.add(Row.of(6, "20.9, 1.8, 1.8"));
df.add(Row.of(7, "21.1, 1.9, 1.8"));
df.add(Row.of(8, "20.9, 2.0, 2.1"));
df.add(Row.of(9, "20.3, 2.3, 2.4"));
df.add(Row.of(10, "22.0, 2.4, 2.5"));
BatchOperator<?> streamData = new MemSourceBatchOp(df, "id int, vec string");
new VectorFunctionBatchOp().setSelectedCol("vec")
.setOutputCol("out").setFuncName("max").linkFrom(streamData).print();
}
}
运行结果
| id | vec | out | | —- | —- | —- |
| 1 | 16.3, 1.1, 1.1 | 16.3 |
| 2 | 16.8, 1.4, 1.5 | 16.8 |
| 3 | 19.2, 1.7, 1.8 | 19.2 |
| 4 | 10.0, 1.7, 1.7 | 10.0 |
| 5 | 19.5, 1.8, 1.9 | 19.5 |
| 6 | 20.9, 1.8, 1.8 | 20.9 |
| 7 | 21.1, 1.9, 1.8 | 21.1 |
| 8 | 20.9, 2.0, 2.1 | 20.9 |
| 9 | 20.3, 2.3, 2.4 | 20.3 |
| 10 | 22.0, 2.4, 2.5 | 22.0 |