GenerateUnsafeProjection生成的代码具有编译错误

Question

以下代码由GenerateUnsafeProjection生成，我将生成的代码粘贴到Java源文件中，但它有编译错误。

import org.apache.spark.sql.catalyst.InternalRow;
import org.apache.spark.sql.catalyst.expressions.UnsafeRow;

public class MyUnsafeProjectionGenerator {
    public java.lang.Object generate(Object[] references) {
        return new SpecificUnsafeProjection(references);
    }
}


class SpecificUnsafeProjection extends org.apache.spark.sql.catalyst.expressions.UnsafeProjection {

    private Object[] references;
    private org.apache.spark.sql.catalyst.expressions.codegen.UnsafeRowWriter[] mutableStateArray_0 = new org.apache.spark.sql.catalyst.expressions.codegen.UnsafeRowWriter[1];

    public SpecificUnsafeProjection(Object[] references) {
        this.references = references;
        mutableStateArray_0[0] = new org.apache.spark.sql.catalyst.expressions.codegen.UnsafeRowWriter(3, 0);

    }

    public void initialize(int partitionIndex) {

    }

//     Scala.Function1 need this
    public java.lang.Object apply(java.lang.Object row) {///Compiling ERROR occurs here
        return apply((InternalRow) row);
    }

    public UnsafeRow apply(InternalRow i) { //
        mutableStateArray_0[0].reset();


        mutableStateArray_0[0].zeroOutNullBytes();

        boolean isNull_0 = i.isNullAt(0);
        int value_0 = isNull_0 ?
                -1 : (i.getInt(0));
        if (isNull_0) {
            mutableStateArray_0[0].setNullAt(0);
        } else {
            mutableStateArray_0[0].write(0, value_0);
        }

        boolean isNull_1 = i.isNullAt(1);
        int value_1 = isNull_1 ?
                -1 : (i.getInt(1));
        if (isNull_1) {
            mutableStateArray_0[0].setNullAt(1);
        } else {
            mutableStateArray_0[0].write(1, value_1);
        }

        boolean isNull_2 = i.isNullAt(3);
        int value_2 = isNull_2 ?
                -1 : (i.getInt(3));
        if (isNull_2) {
            mutableStateArray_0[0].setNullAt(2);
        } else {
            mutableStateArray_0[0].write(2, value_2);
        }
        return (mutableStateArray_0[0].getRow());
    }


}

这两种apply方法都有编译错误。

public UnsafeRow apply(InternalRow i) {没问题，它实现了父类UnsafeProjection中定义的抽象方法

由于UnsafeProjection's parent class Projection extends from InternalRow=>InternalRow是Function1类型，因此生成的代码必须提供另一个应用方法实现：

public java.lang.Object apply(java.lang.Object row) {
    return apply((InternalRow) row);
}

上述申请方法导致编译错误，抱怨

apply(Object) in  org.apache.spark.sql.SpecificUnsafeProjection clashed with apply(T1) in scala.Function1, both has some same erasure,  but neight overrides each other

如果我从

更改apply方法，它将无效

   public java.lang.Object apply(java.lang.Object row) {

到

   public Internal apply(Internal row) {

因为它与其他apply方法冲突：

public UnsafeRow apply(InternalRow i) {

有人可以帮忙解决这个问题吗？