Java 类名：com.alibaba.alink.operator.batch.recommendation.FlattenKObjectBatchOp
Python 类名：FlattenKObjectBatchOp

功能介绍

将推荐结果从json序列化格式转为table格式。

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 | | —- | —- | —- | —- | —- | —- | —- |

| outputCols | 输出结果列列名数组 | 输出结果列列名数组，必选 | String[] | ✓ | | |

| selectedCol | 选中的列名 | 计算列对应的列名 | String | ✓ | 所选列类型为 [STRING] | |

| outputColTypes | 输出结果列列类型数组 | 输出结果列类型数组 | String[] | | | null |

| reservedCols | 算法保留列名 | 算法保留列 | String[] | | | null |

代码示例

Python 代码

  1. from pyalink.alink import *
  3. import pandas as pd
  5. useLocalEnv(1)
  7. df_data = pd.DataFrame([
  8.     [1, 1, 0.6],
  9.     [2, 2, 0.8],
  10.     [2, 3, 0.6],
  11.     [4, 1, 0.6],
  12.     [4, 2, 0.3],
  13.     [4, 3, 0.4],
  14. ])
  16. data = BatchOperator.fromDataframe(df_data, schemaStr='user bigint, item bigint, rating double')
  18. jsonData = Zipped2KObjectBatchOp()\
  19.             .setGroupCol("user")\
  20.             .setObjectCol("item")\
  21.             .setInfoCols(["rating"])\
  22.             .setOutputCol("recomm")\
  23.             .linkFrom(data)\
  24.             .lazyPrint(-1);
  25. recList = FlattenKObjectBatchOp()\
  26.             .setSelectedCol("recomm")\
  27.             .setOutputCols(["item", "rating"])\
  28.             .setOutputColTypes(["long", "double"])\
  29.             .setReservedCols(["user"])\
  30.             .linkFrom(jsonData)\
  31.             .lazyPrint(-1);
  32. BatchOperator.execute();

Java 代码

  1. import org.apache.flink.types.Row;
  3. import com.alibaba.alink.operator.batch.BatchOperator;
  4. import com.alibaba.alink.operator.batch.recommendation.FlattenKObjectBatchOp;
  5. import com.alibaba.alink.operator.batch.source.MemSourceBatchOp;
  6. import com.alibaba.alink.operator.common.recommendation.Zipped2KObjectBatchOp;
  7. import org.junit.Test;
  9. import java.util.Arrays;
  10. import java.util.List;
  12. public class FlattenKObjectBatchOpTest {
  13.     @Test
  14.     public void testFlattenKObjectBatchOp() throws Exception {
  15.         List <Row> df_data = Arrays.asList(
  16.             Row.of(1, 1, 0.6),
  17.             Row.of(2, 2, 0.8),
  18.             Row.of(2, 3, 0.6),
  19.             Row.of(4, 1, 0.6),
  20.             Row.of(4, 2, 0.3),
  21.             Row.of(4, 3, 0.4)
  22.         );
  23.         BatchOperator <?> data = new MemSourceBatchOp(df_data, "user int, item int, rating double");
  24.         BatchOperator <?> jsonData = new Zipped2KObjectBatchOp()
  25.             .setGroupCol("user")
  26.             .setObjectCol("item")
  27.             .setInfoCols("rating")
  28.             .setOutputCol("recomm")
  29.             .linkFrom(data)
  30.             .lazyPrint(-1);
  31.         BatchOperator <?> recList = new FlattenKObjectBatchOp()
  32.             .setSelectedCol("recomm")
  33.             .setOutputCols("item", "rating")
  34.             .setOutputColTypes("long", "double")
  35.             .setReservedCols("user")
  36.             .linkFrom(jsonData)
  37.             .lazyPrint(-1);
  38.         BatchOperator.execute();
  39.     }
  40. }

运行结果

| user | recomm | | —- | —- |

| 1 | {“item”:”[1]”,”rating”:”[0.6]”} |

| 4 | {“item”:”[1,2,3]”,”rating”:”[0.6,0.3,0.4]”} |

| 2 | {“item”:”[2,3]”,”rating”:”[0.8,0.6]”} |

| user | item | rating | | —- | —- | —- |

| 1 | 1 | 0.6000 |

| 4 | 1 | 0.6000 |

| 4 | 2 | 0.3000 |

| 4 | 3 | 0.4000 |

| 2 | 2 | 0.8000 |

| 2 | 3 | 0.6000 |