List of usage examples for org.apache.hadoop.io Text compareTo
@Override public int compareTo(BinaryComparable other)
From source file:boa.datagen.SeqSortMerge.java
License:Apache License
private static int min(Text[] keys) { int index = 0; Text min = keys[0];/*from www . ja v a 2 s. c o m*/ for (int i = 1; i < keys.length; i++) { Text key = keys[i]; if (!key.toString().isEmpty() && key.compareTo(min) < 0) { index = i; min = key; } } return index; }
From source file:cienciaCelularMR.KeyMcell.java
@Override public int compareTo(Object o) { KeyMcell sentiment = (KeyMcell) o;// w w w . j a v a2 s. c o m Text thisValue = this.idUsuario; Text thatValue = sentiment.idUsuario; return this.equals(o) ? 0 : (thatValue.compareTo(thisValue) == 0 ? -1 : 1); }
From source file:com.alexholmes.hadooputils.sort.TextArrayWritable.java
License:Apache License
public int compareTo(TextArrayWritable o) { Writable[] theseTexts = get();/* w w w .jav a2 s .com*/ Writable[] thoseTexts = o.get(); int len = Math.min(theseTexts.length, thoseTexts.length); for (int i = 0; i < len; i++) { Text thisText = (Text) theseTexts[i]; Text thatText = (Text) thoseTexts[i]; int comp = thisText.compareTo(thatText); if (comp != 0) { return comp; } } return 0; }
From source file:com.facebook.presto.accumulo.AccumuloClient.java
License:Apache License
/** * Gets the TabletServer hostname for where the given key is located in the given table * * @param table Fully-qualified table name * @param key Key to locate//from w w w .ja v a2s .c o m * @return The tablet location, or DUMMY_LOCATION if an error occurs */ private Optional<String> getTabletLocation(String table, Key key) { try { // Get the Accumulo table ID so we can scan some fun stuff String tableId = connector.tableOperations().tableIdMap().get(table); // Create our scanner against the metadata table, fetching 'loc' family Scanner scanner = connector.createScanner("accumulo.metadata", auths); scanner.fetchColumnFamily(new Text("loc")); // Set the scan range to just this table, from the table ID to the default tablet // row, which is the last listed tablet Key defaultTabletRow = new Key(tableId + '<'); Key start = new Key(tableId); Key end = defaultTabletRow.followingKey(PartialKey.ROW); scanner.setRange(new Range(start, end)); Optional<String> location = Optional.empty(); if (key == null) { // if the key is null, then it is -inf, so get first tablet location Iterator<Entry<Key, Value>> iter = scanner.iterator(); if (iter.hasNext()) { location = Optional.of(iter.next().getValue().toString()); } } else { // Else, we will need to scan through the tablet location data and find the location // Create some text objects to do comparison for what we are looking for Text splitCompareKey = new Text(); key.getRow(splitCompareKey); Text scannedCompareKey = new Text(); // Scan the table! for (Entry<Key, Value> entry : scanner) { // Get the bytes of the key byte[] keyBytes = entry.getKey().getRow().copyBytes(); // If the last byte is <, then we have hit the default tablet, so use this location if (keyBytes[keyBytes.length - 1] == '<') { location = Optional.of(entry.getValue().toString()); break; } else { // Chop off some magic nonsense scannedCompareKey.set(keyBytes, 3, keyBytes.length - 3); // Compare the keys, moving along the tablets until the location is found if (scannedCompareKey.getLength() > 0) { int compareTo = splitCompareKey.compareTo(scannedCompareKey); if (compareTo <= 0) { location = Optional.of(entry.getValue().toString()); } else { // all future tablets will be greater than this key break; } } } } scanner.close(); } // If we were unable to find the location for some reason, return the default tablet // location return location.isPresent() ? location : getDefaultTabletLocation(table); } catch (Exception e) { // Swallow this exception so the query does not fail due to being unable // to locate the tablet server for the provided Key. // This is purely an optimization, but we will want to log the error. LOG.error("Failed to get tablet location, returning dummy location", e); return Optional.empty(); } }
From source file:com.inmobi.conduit.distcp.tools.mapred.CopyCommitter.java
License:Apache License
private void deleteMissing(Configuration conf) throws IOException { LOG.info("-delete option is enabled. About to remove entries from " + "target that are missing in source"); Path sourceListing = new Path(conf.get(DistCpConstants.CONF_LABEL_LISTING_FILE_PATH)); FileSystem clusterFS = sourceListing.getFileSystem(conf); Path sortedSourceListing = DistCpUtils.sortListing(clusterFS, conf, sourceListing); Path targetListing = new Path(sourceListing.getParent(), "targetListing.seq"); CopyListing target = new GlobbedCopyListing(conf, null); List<Path> targets = new ArrayList<Path>(1); Path targetFinalPath = new Path(conf.get(DistCpConstants.CONF_LABEL_TARGET_FINAL_PATH)); targets.add(targetFinalPath);/* w w w.j ava2 s . c o m*/ DistCpOptions options = new DistCpOptions(targets, new Path("/NONE")); target.buildListing(targetListing, options); Path sortedTargetListing = DistCpUtils.sortListing(clusterFS, conf, targetListing); long totalLen = clusterFS.getFileStatus(sortedTargetListing).getLen(); SequenceFile.Reader sourceReader = new SequenceFile.Reader(clusterFS, sortedSourceListing, conf); SequenceFile.Reader targetReader = new SequenceFile.Reader(clusterFS, sortedTargetListing, conf); long deletedEntries = 0; try { FileStatus srcFileStatus = new FileStatus(); Text srcRelPath = new Text(); FileStatus trgtFileStatus = new FileStatus(); Text trgtRelPath = new Text(); FileSystem targetFS = targetFinalPath.getFileSystem(conf); boolean srcAvailable = sourceReader.next(srcRelPath, srcFileStatus); while (targetReader.next(trgtRelPath, trgtFileStatus)) { while (srcAvailable && trgtRelPath.compareTo(srcRelPath) > 0) { srcAvailable = sourceReader.next(srcRelPath, srcFileStatus); } if (srcAvailable && trgtRelPath.equals(srcRelPath)) continue; boolean result = (!targetFS.exists(trgtFileStatus.getPath()) || targetFS.delete(trgtFileStatus.getPath(), true)); if (result) { LOG.info("Deleted " + trgtFileStatus.getPath() + " - Missing at source"); deletedEntries++; } else { throw new IOException("Unable to delete " + trgtFileStatus.getPath()); } HadoopCompat.progress(taskAttemptContext); HadoopCompat.setStatus(taskAttemptContext, "Deleting missing files from target. [" + targetReader.getPosition() * 100 / totalLen + "%]"); } } finally { IOUtils.closeStream(sourceReader); IOUtils.closeStream(targetReader); } LOG.info("Deleted " + deletedEntries + " from target: " + targets.get(0)); }
From source file:com.kylinolap.job.hadoop.cube.RowKeyDistributionCheckerMapper.java
License:Apache License
@Override public void map(Text key, Text value, Context context) throws IOException, InterruptedException { for (Text t : keyList) { if (key.compareTo(t) < 0) { Long v = resultMap.get(t); long length = key.getLength() + value.getLength(); v += length;/* w w w .j av a 2s . co m*/ resultMap.put(t, v); break; } } }
From source file:com.marcolotz.MRComponents.KeyStructureWritable.java
License:Creative Commons License
@Override public int compareTo(KeyStructureWritable comparedKeyStruct) { Text seriesInstanceHere = new Text(this.getSeriesInstanceUID()); Text seriesInstanceCompared = new Text(comparedKeyStruct.getSeriesInstanceUID()); return seriesInstanceHere.compareTo(seriesInstanceCompared); }
From source file:com.pinterest.hdfsbackup.distcp.DistCp.java
License:Apache License
/** Delete the dst files/dirs which do not exist in src */ static private void deleteNonexisting(FileSystem dstfs, FileStatus dstroot, Path dstsorted, FileSystem jobfs, Path jobdir, JobConf jobconf, Configuration conf) throws IOException { if (!dstroot.isDir()) { throw new IOException("dst must be a directory when option " + Options.DELETE.cmd + " is set, but dst (= " + dstroot.getPath() + ") is not a directory."); }//from w w w. j a va2s . c o m //write dst lsr results final Path dstlsr = new Path(jobdir, "_distcp_dst_lsr"); final SequenceFile.Writer writer = SequenceFile.createWriter(jobfs, jobconf, dstlsr, Text.class, FileStatus.class, SequenceFile.CompressionType.NONE); try { //do lsr to get all file statuses in dstroot final Stack<FileStatus> lsrstack = new Stack<FileStatus>(); for (lsrstack.push(dstroot); !lsrstack.isEmpty();) { final FileStatus status = lsrstack.pop(); if (status.isDir()) { for (FileStatus child : dstfs.listStatus(status.getPath())) { String relative = makeRelative(dstroot.getPath(), child.getPath()); writer.append(new Text(relative), child); lsrstack.push(child); } } } } finally { checkAndClose(writer); } //sort lsr results final Path sortedlsr = new Path(jobdir, "_distcp_dst_lsr_sorted"); SequenceFile.Sorter sorter = new SequenceFile.Sorter(jobfs, new Text.Comparator(), Text.class, FileStatus.class, jobconf); sorter.sort(dstlsr, sortedlsr); //compare lsr list and dst list SequenceFile.Reader lsrin = null; SequenceFile.Reader dstin = null; try { lsrin = new SequenceFile.Reader(jobfs, sortedlsr, jobconf); dstin = new SequenceFile.Reader(jobfs, dstsorted, jobconf); //compare sorted lsr list and sorted dst list final Text lsrpath = new Text(); final FileStatus lsrstatus = new FileStatus(); final Text dstpath = new Text(); final Text dstfrom = new Text(); final FsShell shell = new FsShell(conf); final String[] shellargs = { "-rmr", null }; boolean hasnext = dstin.next(dstpath, dstfrom); for (; lsrin.next(lsrpath, lsrstatus);) { int dst_cmp_lsr = dstpath.compareTo(lsrpath); for (; hasnext && dst_cmp_lsr < 0;) { hasnext = dstin.next(dstpath, dstfrom); dst_cmp_lsr = dstpath.compareTo(lsrpath); } if (dst_cmp_lsr == 0) { //lsrpath exists in dst, skip it hasnext = dstin.next(dstpath, dstfrom); } else { //lsrpath does not exist, delete it String s = new Path(dstroot.getPath(), lsrpath.toString()).toString(); if (shellargs[1] == null || !isAncestorPath(shellargs[1], s)) { shellargs[1] = s; int r = 0; try { r = shell.run(shellargs); } catch (Exception e) { throw new IOException("Exception from shell.", e); } if (r != 0) { throw new IOException( "\"" + shellargs[0] + " " + shellargs[1] + "\" returns non-zero value " + r); } } } } } finally { checkAndClose(lsrin); checkAndClose(dstin); } }
From source file:com.scaleunlimited.cascading.DistCp.java
License:Apache License
/** Delete the dst files/dirs which do not exist in src */ static private void deleteNonexisting(FileSystem dstfs, FileStatus dstroot, Path dstsorted, FileSystem jobfs, Path jobdir, JobConf jobconf, Configuration conf) throws IOException { if (!dstroot.isDir()) { throw new IOException("dst must be a directory when option " + Options.DELETE.cmd + " is set, but dst (= " + dstroot.getPath() + ") is not a directory."); }// ww w . j a v a 2 s .com //write dst lsr results final Path dstlsr = new Path(jobdir, "_distcp_dst_lsr"); final SequenceFile.Writer writer = SequenceFile.createWriter(jobfs, jobconf, dstlsr, Text.class, FileStatus.class, SequenceFile.CompressionType.NONE); try { //do lsr to get all file statuses in dstroot final Stack<FileStatus> lsrstack = new Stack<FileStatus>(); for (lsrstack.push(dstroot); !lsrstack.isEmpty();) { final FileStatus status = lsrstack.pop(); if (status.isDir()) { for (FileStatus child : dstfs.listStatus(status.getPath())) { String relative = makeRelative(dstroot.getPath(), child.getPath()); writer.append(new Text(relative), child); lsrstack.push(child); } } } } finally { checkAndClose(writer); } //sort lsr results final Path sortedlsr = new Path(jobdir, "_distcp_dst_lsr_sorted"); SequenceFile.Sorter sorter = new SequenceFile.Sorter(jobfs, new Text.Comparator(), Text.class, FileStatus.class, jobconf); sorter.sort(dstlsr, sortedlsr); //compare lsr list and dst list SequenceFile.Reader lsrin = null; SequenceFile.Reader dstin = null; try { lsrin = new SequenceFile.Reader(jobfs, sortedlsr, jobconf); dstin = new SequenceFile.Reader(jobfs, dstsorted, jobconf); //compare sorted lsr list and sorted dst list final Text lsrpath = new Text(); final FileStatus lsrstatus = new FileStatus(); final Text dstpath = new Text(); final Text dstfrom = new Text(); final FsShell shell = new FsShell(conf); final String[] shellargs = { "-rmr", null }; boolean hasnext = dstin.next(dstpath, dstfrom); for (; lsrin.next(lsrpath, lsrstatus);) { int dst_cmp_lsr = dstpath.compareTo(lsrpath); for (; hasnext && dst_cmp_lsr < 0;) { hasnext = dstin.next(dstpath, dstfrom); dst_cmp_lsr = dstpath.compareTo(lsrpath); } if (dst_cmp_lsr == 0) { //lsrpath exists in dst, skip it hasnext = dstin.next(dstpath, dstfrom); } else { //lsrpath does not exist, delete it String s = new Path(dstroot.getPath(), lsrpath.toString()).toString(); if (shellargs[1] == null || !isAncestorPath(shellargs[1], s)) { shellargs[1] = s; int r = 0; try { r = shell.run(shellargs); } catch (Exception e) { throw new IOException("Exception from shell.", e); } if (r != 0) { throw new IOException( "\"" + shellargs[0] + " " + shellargs[1] + "\" returns non-zero value " + r); } } } } } finally { checkAndClose(lsrin); checkAndClose(dstin); } }
From source file:fr.ens.biologie.genomique.eoulsan.modules.mgmt.hadoop.DistCp.java
License:LGPL
/** Delete the dst files/dirs which do not exist in src */ static private void deleteNonexisting(final FileSystem dstfs, final FileStatus dstroot, final Path dstsorted, final FileSystem jobfs, final Path jobdir, final JobConf jobconf, final Configuration conf) throws IOException { if (!dstroot.isDir()) { throw new IOException("dst must be a directory when option " + Options.DELETE.cmd + " is set, but dst (= " + dstroot.getPath() + ") is not a directory."); }/*from ww w.java2 s .com*/ // write dst lsr results final Path dstlsr = new Path(jobdir, "_distcp_dst_lsr"); final SequenceFile.Writer writer = SequenceFile.createWriter(jobfs, jobconf, dstlsr, Text.class, dstroot.getClass(), SequenceFile.CompressionType.NONE); try { // do lsr to get all file statuses in dstroot final Stack<FileStatus> lsrstack = new Stack<>(); for (lsrstack.push(dstroot); !lsrstack.isEmpty();) { final FileStatus status = lsrstack.pop(); if (status.isDir()) { for (FileStatus child : dstfs.listStatus(status.getPath())) { String relative = makeRelative(dstroot.getPath(), child.getPath()); writer.append(new Text(relative), child); lsrstack.push(child); } } } } finally { checkAndClose(writer); } // sort lsr results final Path sortedlsr = new Path(jobdir, "_distcp_dst_lsr_sorted"); SequenceFile.Sorter sorter = new SequenceFile.Sorter(jobfs, new Text.Comparator(), Text.class, FileStatus.class, jobconf); sorter.sort(dstlsr, sortedlsr); // compare lsr list and dst list SequenceFile.Reader lsrin = null; SequenceFile.Reader dstin = null; try { lsrin = new SequenceFile.Reader(jobfs, sortedlsr, jobconf); dstin = new SequenceFile.Reader(jobfs, dstsorted, jobconf); // compare sorted lsr list and sorted dst list final Text lsrpath = new Text(); final FileStatus lsrstatus = new FileStatus(); final Text dstpath = new Text(); final Text dstfrom = new Text(); final FsShell shell = new FsShell(conf); final String[] shellargs = { "-rmr", null }; boolean hasnext = dstin.next(dstpath, dstfrom); for (; lsrin.next(lsrpath, lsrstatus);) { int dst_cmp_lsr = dstpath.compareTo(lsrpath); for (; hasnext && dst_cmp_lsr < 0;) { hasnext = dstin.next(dstpath, dstfrom); dst_cmp_lsr = dstpath.compareTo(lsrpath); } if (dst_cmp_lsr == 0) { // lsrpath exists in dst, skip it hasnext = dstin.next(dstpath, dstfrom); } else { // lsrpath does not exist, delete it String s = new Path(dstroot.getPath(), lsrpath.toString()).toString(); if (shellargs[1] == null || !isAncestorPath(shellargs[1], s)) { shellargs[1] = s; int r = 0; try { r = shell.run(shellargs); } catch (Exception e) { throw new IOException("Exception from shell.", e); } if (r != 0) { throw new IOException( "\"" + shellargs[0] + " " + shellargs[1] + "\" returns non-zero value " + r); } } } } } finally { checkAndClose(lsrin); checkAndClose(dstin); } }