本文整理汇总了Java中org.apache.hadoop.mapred.Merger.Segment.getActualPosition方法的典型用法代码示例。如果您正苦于以下问题：Java Segment.getActualPosition方法的具体用法？Java Segment.getActualPosition怎么用？Java Segment.getActualPosition使用的例子？那么, 这里精选的方法代码示例或许可以为您提供帮助。您也可以进一步了解该方法所在类org.apache.hadoop.mapred.Merger.Segment的用法示例。
在下文中一共展示了Segment.getActualPosition方法的1个代码示例，这些例子默认根据受欢迎程度排序。您可以为喜欢或者感觉有用的代码点赞，您的评价将有助于系统推荐出更棒的Java代码示例。
示例1: hasNext
import org.apache.hadoop.mapred.Merger.Segment; //导入方法依赖的package包/类
public boolean hasNext() throws IOException {
if (lastSegmentEOF) {
return false;
}
// We read the next KV from the cache to decide if there is any left.
// Since hasNext can be called several times before the actual call to
// next(), we use hasMore to avoid extra reads. hasMore is set to false
// when the user actually consumes this record in next()
if (hasMore) {
return true;
}
Segment<K,V> seg = segmentList.get(readSegmentIndex);
// Mark the current position. This would be set to currentKVOffset
// when the user consumes this record in next().
nextKVOffset = (int) seg.getActualPosition();
if (seg.nextRawKey()) {
currentKey = seg.getKey();
seg.getValue(currentValue);
hasMore = true;
return true;
} else {
if (!seg.inMemory()) {
seg.closeReader();
}
}
// If this is the last segment, mark the lastSegmentEOF flag and return
if (readSegmentIndex == segmentList.size() - 1) {
nextKVOffset = -1;
lastSegmentEOF = true;
return false;
}
nextKVOffset = 0;
readSegmentIndex ++;
Segment<K,V> nextSegment = segmentList.get(readSegmentIndex);
// We possibly are moving from a memory segment to a disk segment.
// Reset so that we do not corrupt the in-memory segment buffer.
// See HADOOP-5494
if (!nextSegment.inMemory()) {
currentValue.reset(currentDiskValue.getData(),
currentDiskValue.getLength());
nextSegment.init(null);
}
if (nextSegment.nextRawKey()) {
currentKey = nextSegment.getKey();
nextSegment.getValue(currentValue);
hasMore = true;
return true;
} else {
throw new IOException("New segment did not have even one K/V");
}
}