public void map(LongWritable key, Text value, Context context)
throws IOException, InterruptedException {
String valstr = value.toString().trim();
if (valstr.length() == 0) return;
DataPoint dp = new DataPoint(valstr);
double[] vals = dp.getFeatures().getValArr();
if (featnum != vals.length) throw new RuntimeException("当前record的特征长度与设置长度不同！");
for (int i = 0; i < vals.length; i++) {
context.write(new Text(String.valueOf(i)), new Text(String.valueOf(vals[i])));
}
}
protected void map(LongWritable key, Text value, Context context)
throws IOException, InterruptedException {
String rowkey_prefix = filename + "," + key.get();
// 输入数据的格式是 ： uid\tbid\tlabel\tweigth\tweigth\tweigth\tweigth\tweigth\tweigth\t
if (value.toString().split("\\s+").length < 4) return;
DataPoint point = new DataPoint(value.toString());
if (point.getFeatures() == null) return;
ArrayList<Double> features = point.getFeatures().getValue();
int slicenum = 0;
ArrayList<Double> slice = new ArrayList<Double>(slicelen);
for (int i = 0; i < features.size(); i++) {
int index = i - slicenum * slicelen; // sub-feature index
if (index < slicelen) {
double cur_feat = features.get(i);
slice.add(cur_feat);
} else {
// 注意到 这里的arraylist 用的是deepcopy ， 长度满足sliceLen 就输出
WDataPoint sub_dp =
new WDataPoint(
point.getUid(), point.getBid(), point.getLabel(), slice, 0.0, slicenum);
context.write(new Text(rowkey_prefix + "," + slicenum), new Text(sub_dp.toString()));
// 降档前特征加入下一条子特征数组内
slicenum++;
slice.clear(); // 可以清空， 最底层实现了安置复制
double cur_feat = features.get(i);
slice.add(cur_feat); // 这时sub_index == 0
}
}
if (slice.size() != 0) {
WDataPoint sub_dp =
new WDataPoint(point.getUid(), point.getBid(), point.getLabel(), slice, 0.0, slicenum);
context.write(new Text(rowkey_prefix + "," + slicenum), new Text(sub_dp.toString()));
slice.clear(); // 可以清空， 最底层实现了安置复制
}
}
