@Override public FileInputSplit[] createInputSplits(int minNumSplits) throws IOException { List<FileStatus> files = this.getFiles(); final FileSystem fs = this.filePath.getFileSystem(); final long blockSize = this.blockSize == NATIVE_BLOCK_SIZE ? fs.getDefaultBlockSize() : this.blockSize; final List<FileInputSplit> inputSplits = new ArrayList<FileInputSplit>(minNumSplits); for (FileStatus file : files) { long splitSize = blockSize; for (long pos = 0, length = file.getLen(); pos < length; pos += splitSize) { long remainingLength = Math.min(pos + splitSize, length) - pos; // get the block locations and make sure they are in order with respect to their offset final BlockLocation[] blocks = fs.getFileBlockLocations(file, pos, remainingLength); Arrays.sort(blocks); inputSplits.add(new FileInputSplit(inputSplits.size(), file.getPath(), pos, remainingLength, blocks[0].getHosts())); } } if (inputSplits.size() < minNumSplits) { LOG.warn(String.format( "With the given block size %d, the file %s cannot be split into %d blocks. Filling up with empty splits...", blockSize, this.filePath, minNumSplits)); FileStatus last = files.get(files.size() - 1); final BlockLocation[] blocks = fs.getFileBlockLocations(last, 0, last.getLen()); for (int index = files.size(); index < minNumSplits; index++) { inputSplits.add(new FileInputSplit(index, last.getPath(), last.getLen(), 0, blocks[0].getHosts())); } } return inputSplits.toArray(new FileInputSplit[0]); }
@Override public FileInputSplit[] createInputSplits(final int minNumSplits) throws IOException { final List<FileStatus> files = this.getFiles(); final FileSystem fs = this.filePath.getFileSystem(); final long blockSize = this.blockSize == NATIVE_BLOCK_SIZE ? fs.getDefaultBlockSize() : this.blockSize; final List<FileInputSplit> inputSplits = new ArrayList<FileInputSplit>(minNumSplits); for (final FileStatus file : files) { final long splitSize = blockSize; for (long pos = 0, length = file.getLen(); pos < length; pos += splitSize) { final long remainingLength = Math.min(pos + splitSize, length) - pos; // get the block locations and make sure they are in order with respect to their offset final BlockLocation[] blocks = fs.getFileBlockLocations(file, pos, remainingLength); Arrays.sort(blocks); inputSplits.add(new FileInputSplit(inputSplits.size(), file.getPath(), pos, remainingLength, blocks[0].getHosts())); } } if (inputSplits.size() < minNumSplits) { LOG.warn(String.format( "With the given block size %d, the file %s cannot be split into %d blocks. Filling up with empty splits...", blockSize, this.filePath, minNumSplits)); final FileStatus last = files.get(files.size() - 1); final BlockLocation[] blocks = fs.getFileBlockLocations(last, 0, last.getLen()); for (int index = files.size(); index < minNumSplits; index++) inputSplits.add(new FileInputSplit(index, last.getPath(), last.getLen(), 0, blocks[0].getHosts())); } return inputSplits.toArray(new FileInputSplit[0]); }
@Override public FileInputSplit[] createInputSplits(int minNumSplits) throws IOException { List<FileStatus> files = this.getFiles(); final FileSystem fs = this.filePath.getFileSystem(); final long blockSize = this.blockSize == NATIVE_BLOCK_SIZE ? fs.getDefaultBlockSize() : this.blockSize; final List<FileInputSplit> inputSplits = new ArrayList<FileInputSplit>(minNumSplits); for (FileStatus file : files) { long splitSize = blockSize; for (long pos = 0, length = file.getLen(); pos < length; pos += splitSize) { long remainingLength = Math.min(pos + splitSize, length) - pos; // get the block locations and make sure they are in order with respect to their offset final BlockLocation[] blocks = fs.getFileBlockLocations(file, pos, remainingLength); Arrays.sort(blocks); inputSplits.add(new FileInputSplit(inputSplits.size(), file.getPath(), pos, remainingLength, blocks[0].getHosts())); } } if (inputSplits.size() < minNumSplits) { LOG.warn(String.format( "With the given block size %d, the file %s cannot be split into %d blocks. Filling up with empty splits...", blockSize, this.filePath, minNumSplits)); FileStatus last = files.get(files.size() - 1); final BlockLocation[] blocks = fs.getFileBlockLocations(last, 0, last.getLen()); for (int index = files.size(); index < minNumSplits; index++) { inputSplits.add(new FileInputSplit(index, last.getPath(), last.getLen(), 0, blocks[0].getHosts())); } } return inputSplits.toArray(new FileInputSplit[0]); }
int splitNum = 0; for (final FileStatus file : files) { final BlockLocation[] blocks = fs.getFileBlockLocations(file, 0, file.getLen()); Set<String> hosts = new HashSet<String>(); for(BlockLocation block : blocks) { final BlockLocation[] blocks = fs.getFileBlockLocations(file, 0, len); Arrays.sort(blocks); final BlockLocation[] blocks = fs.getFileBlockLocations(file, 0, 0); String[] hosts; if (blocks.length > 0) {
int splitNum = 0; for (final FileStatus file : files) { final BlockLocation[] blocks = fs.getFileBlockLocations(file, 0, file.getLen()); Set<String> hosts = new HashSet<String>(); for(BlockLocation block : blocks) { final BlockLocation[] blocks = fs.getFileBlockLocations(file, 0, len); Arrays.sort(blocks); final BlockLocation[] blocks = fs.getFileBlockLocations(file, 0, 0); String[] hosts; if (blocks.length > 0) {
final BlockLocation[] blocks = fs.getFileBlockLocations(file, 0, len); Arrays.sort(blocks); final BlockLocation[] blocks = fs.getFileBlockLocations(file, 0, 0); String[] hosts; if (blocks.length > 0) {
final BlockLocation[] blocks = fs.getFileBlockLocations(file, 0, len); Arrays.sort(blocks); final BlockLocation[] blocks = fs.getFileBlockLocations(file, 0, 0); String[] hosts; if (blocks.length > 0) {
FileStatus fileStatus = fs.getFileStatus(normalizedPath); BlockLocation[] blocks = fs.getFileBlockLocations(fileStatus, 0, fileStatus.getLen()); inputFormat.open(new FileInputSplit(0, new Path(path), 0, fileStatus.getLen(), blocks[0].getHosts())); return inputFormat;
FileStatus fileStatus = fs.getFileStatus(normalizedPath); BlockLocation[] blocks = fs.getFileBlockLocations(fileStatus, 0, fileStatus.getLen()); inputFormat.open(new FileInputSplit(0, new Path(path), 0, fileStatus.getLen(), blocks[0].getHosts())); return inputFormat;