IForestOutlierBatchOp进行异常检测时，输出结果顺序颠倒

`

String[] FEATURE_NAMES = {"f1", "f2", "f3", "f4", "f5",
"f6", "f7", "f8", "f9"};
String[] ALL_NAMES = {"f1", "f2", "f3", "f4", "f5",
"f6", "f7", "f8", "f9", "label"};
BatchOperator data = new MemSourceBatchOp(
new Object[][] {
{10, 10, 10, 10, 10, 10, 10, 10, 10, 0},
{10, 10, 10, 10, 10, 10, 10, 10, 10, 0},
{10, 10, 10, 10, 10, 10, 10, 10, 10, 0},
{10, 10, 10, 10, 10, 10, 10, 10, 10, 0},
{10, 10, 10, 10, 10, 10, 10, 10, 10, 0},
{1, 1, 1, 1, 1, 1, 1, 1, 100, 1},
{1, 1, 1, 1, 1, 1, 1, 1, 200, 1},
{1, 1, 1, 1, 1, 1, 1, 1, 300, 1},
},
ALL_NAMES);
BatchOperator outlier = new IForestOutlierBatchOp()
.setFeatureCols(FEATURE_NAMES)
.setPredictionCol("pred")
.setPredictionDetailCol("pred_detail");
data.link(outlier).print();

`
的运行结果为：

f1 |f2 |f3 |f4 |f5 |f6 |f7 |f8 |f9 |label|pred|pred_detail
---|---|---|---|---|---|---|---|---|-----|----|-----------
10|10|10|10|10|10|10|10|10|0|true|{"outlier_score":"0.5907573613555553","is_outlier":"true"}
10|10|10|10|10|10|10|10|10|0|true|{"outlier_score":"0.5496377996962295","is_outlier":"true"}
10|10|10|10|10|10|10|10|10|0|true|{"outlier_score":"0.5988412960965512","is_outlier":"true"}
10|10|10|10|10|10|10|10|10|0|false|{"outlier_score":"0.48774601295119835","is_outlier":"false"}
10|10|10|10|10|10|10|10|10|0|false|{"outlier_score":"0.48774601295119835","is_outlier":"false"}
1|1|1|1|1|1|1|1|100|1|false|{"outlier_score":"0.48774601295119835","is_outlier":"false"}
1|1|1|1|1|1|1|1|200|1|false|{"outlier_score":"0.48774601295119835","is_outlier":"false"}
1|1|1|1|1|1|1|1|300|1|false|{"outlier_score":"0.48774601295119835","is_outlier":"false"}

后三行的结果跑到前三行中

Start at the IForestOutlierBatchOp entry point and reproduce the ordering problem with the Java example and data shown in the issue. Trace how prediction results are joined back to the input rows, then verify that pred and pred_detail remain aligned with their original rows and run the relevant existing tests, if present.

Written by the indexing model from the issue text.