Java ArrayIndexOutOfBoundsException错误

Java ArrayIndexOutOfBoundsException错误,java,hadoop,Java,Hadoop,当我运行代码时,我从reducer的任务中得到ArrayIndexOutOfBoundsException错误 我的代码如下: public void map(ImageHeader key, FloatImage value, Context context) throws IOException, InterruptedException{ if (value != null) { mapcounter++; FloatImage gray=new F

当我运行代码时,我从reducer的任务中得到
ArrayIndexOutOfBoundsException
错误

我的代码如下:

public void map(ImageHeader key, FloatImage value, Context context) throws IOException, InterruptedException{
    if (value != null) {
        mapcounter++;
        FloatImage gray=new FloatImage(value.getWidth(),value.getHeight(),value.getBands());

        int imageWidth = value.getWidth();
        int imageHeight = value.getHeight();

        for (int x = 0; x < imageWidth-1; x++) {

            for (int y = 0; y < imageHeight-1; y++) {
                float red =value.getPixel(x, y, 0);
                float green =value.getPixel(x, y, 1);
                float blue =value.getPixel(x, y, 2);
                //average of RGB
                float avg = (red + blue + green)/3;

                //set R, G & B with avg color
                gray.setPixel(x, y, 0, avg);
                gray.setPixel(x, y, 1, avg);
                gray.setPixel(x, y, 2, avg);
            }
        }

        ImageEncoder encoder = JPEGImageUtil.getInstance();

        FSDataOutputStream os = fileSystem.create(outpath);
        encoder.encodeImage(gray, key, os);
        os.flush();
        os.close();

        context.write(new BooleanWritable(true), new LongWritable(1));
    }
    else
        context.write(new BooleanWritable(false), new LongWritable(0));
}
public static class MyReducer extends Reducer<BooleanWritable, LongWritable, BooleanWritable, LongWritable> {

    public void reduce(BooleanWritable key, Iterable<LongWritable> values, Context context)
        throws IOException, InterruptedException
        {
            System.out.println("REDUCING");
            for (LongWritable temp_hash : values)
            {
                   context.write(new BooleanWritable(true), new LongWritable(1));
            }//for
        }
}
我如何解决这个问题


第二个问题:如何在程序中忽略reduce阶段而不运行reduce阶段?

看起来这是Apache hoop中的一个bug。他们声称他们已修复了以下内容

它已作为MAPREDUCE-365的一部分修复


有关该错误的详细信息,请参阅。

谢谢Jayamohan。我查过了。成功了。非常感谢:)