Java Code Examples for org.apache.accumulo.core.data.Key#getColumnFamily()

@Override
public void seek(Range range, Collection<ByteSequence> columnFamilies, boolean inclusive) throws IOException {
    // save parameters for future internal seeks
    latestRange = range;
    latestColumnFamilies = columnFamilies;
    latestInclusive = inclusive;
    lastRowFound = null;

    Key startKey = range.getStartKey();
    Range seekRange = new Range(startKey == null ? null : new Key(startKey.getRow(), startKey.getColumnFamily()), true, range.getEndKey(), range.isEndKeyInclusive());
    super.seek(seekRange, columnFamilies, inclusive);
    finished = false;

    if (getSource().hasTop()) {
        lastRowFound = getSource().getTopKey().getRow();
        if (range.beforeStartKey(getSource().getTopKey()))
            consume();
    }
}
 

protected boolean isEventKey(Key k) {
    Text cf = k.getColumnFamily();
    return cf.getLength() > 0
                    && cf.find("\u0000") != -1
                    && !((cf.charAt(0) == 'f' && cf.charAt(1) == 'i' && cf.charAt(2) == 0) || (cf.getLength() == 1 && cf.charAt(0) == 'd') || (cf
                                    .getLength() == 2 && cf.charAt(0) == 't' && cf.charAt(1) == 'f'));
}
 

@Override
public Key getTopKey() {
    Key topKey = super.getTopKey();

    long timestamp = reverseTimestampToNormalTime(Long.parseLong(topKey.getRow().toString()));

    Key retKey = new Key(new Text(truncatedReverseTimestamp(timestamp, bucketSize).toString()),
            topKey.getColumnFamily(), topKey.getColumnQualifier(),
            new Text(topKey.getColumnVisibility().toString()), topKey.getTimestamp());

    return retKey;
}
 

@Override
public int getPartition(BulkIngestKey bKey, Value value, int reducers) {
    HashCodeBuilder hcb = new HashCodeBuilder(157, 41);
    Key key = bKey.getKey();
    Text cf = key.getColumnFamily();
    if (colFams.contains(cf)) {
        hcb.append(cf);
    }
    hcb.append(key.getRow());
    int partition = (hcb.toHashCode() >>> 1) % reducers;
    if (log.isTraceEnabled()) {
        log.trace("Returning " + partition + " for BIK " + bKey);
    }
    return partition;
}
 

/**
 * Checks to see if the provided term appears in other documents.
 * @param term the term to search for.
 * @param currentDocId the current document ID that the search term exists in.
 * @return {@code true} if the term was found in other documents. {@code false} otherwise.
 */
private boolean doesTermExistInOtherDocs(final String term, final int currentDocId, final Text docIdText) {
    try {
        final String freeTextDocTableName = getFreeTextDocTablename(conf);
        final Scanner scanner = getScanner(freeTextDocTableName);

        final String t = StringUtils.removeEnd(term, "*").toLowerCase();
        final Text queryTerm = ColumnPrefixes.getTermColFam(t);

        // perform query and read results
        scanner.fetchColumnFamily(queryTerm);

        for (final Entry<Key, Value> entry : scanner) {
            final Key key = entry.getKey();
            final Text row = key.getRow();
            final int rowId = Integer.parseInt(row.toString());
            // We only want to check other documents from the one we're deleting
            if (rowId != currentDocId) {
                final Text columnFamily = key.getColumnFamily();
                final String columnFamilyValue = columnFamily.toString();
                // Check that the value has the term prefix
                if (columnFamilyValue.startsWith(ColumnPrefixes.TERM_CF_PREFIX.toString())) {
                    final Text text = ColumnPrefixes.removePrefix(columnFamily);
                    final String value = text.toString();
                    if (value.equals(term)) {
                        return true;
                    }
                }
            }
        }
    } catch (final IOException e) {
        logger.error("Error searching for the existance of the term in other documents", e);
    }
    return false;
}
 

@Override
protected Key getSeekStartKey(Key current, ByteSequence pointer) {
    int lastNullIndex = current.getColumnQualifier().toString().lastIndexOf(Constants.NULL);
    lastNullIndex = current.getColumnQualifier().toString().lastIndexOf(Constants.NULL, lastNullIndex - 1);
    String prefix = current.getColumnQualifier().toString().substring(0, lastNullIndex + 1);
    return new Key(current.getRow(), current.getColumnFamily(), new Text(prefix + pointer + Constants.MAX_UNICODE_STRING));
}
 

private Key newDocumentKey(final Key fieldKey, long timestamp) {
    final Text row = fieldKey.getRow();
    final Text cf = fieldKey.getColumnFamily();
    final Text cq = new Text();
    final Text visibility = new Text();
    return new Key(row, cf, cq, visibility, timestamp);
}
 

private Text constructColFamilyForResult(Key lastSeenKey, String lastSeenDate) {
    return lastSeenKey.getColumnFamily();
}
 

public ValueComparator(Key metadata) {
    fi = (metadata == null ? new Text() : metadata.getColumnFamily());
}
 

@Test
public void testLotsOfCollisions() throws Exception {

  Random rand = new Random(45734985);

  int[] nums = new int[1000];
  int[] expectedCounts = new int[5];

  for (int i = 0; i < nums.length; i++) {
    nums[i] = rand.nextInt(expectedCounts.length);
    expectedCounts[nums[i]]++;
  }

  try (LoaderExecutor loader = client.newLoaderExecutor()) {
    for (int num : nums) {
      loader.execute(new NumLoader(num));
    }
  }

  miniFluo.waitForObservers();

  long recentTS;

  try (Snapshot snapshot = client.newSnapshot()) {

    for (int i = 0; i < expectedCounts.length; i++) {
      String total = snapshot.gets(i + "", STAT_TOTAL);
      Assert.assertNotNull(total);
      Assert.assertEquals(expectedCounts[i], Integer.parseInt(total));
      String processed = snapshot.gets(i + "", STAT_PROCESSED);
      Assert.assertNotNull(processed);
      Assert.assertEquals(expectedCounts[i], Integer.parseInt(processed));
    }

    String allTotal = snapshot.gets("all", STAT_TOTAL);
    Assert.assertNotNull(allTotal);
    Assert.assertEquals(1000, Integer.parseInt(allTotal));

    recentTS = snapshot.getStartTimestamp();
  }

  long oldestTS = ZookeeperUtil.getGcTimestamp(config.getAppZookeepers());

  while (oldestTS < recentTS) {
    UtilWaitThread.sleep(300);
    oldestTS = ZookeeperUtil.getGcTimestamp(config.getAppZookeepers());
  }

  aClient.tableOperations().compact(getCurTableName(), null, null, true, true);

  Scanner scanner = aClient.createScanner(getCurTableName(), Authorizations.EMPTY);

  HashSet<String> rowCols = new HashSet<>();

  for (Entry<Key, Value> entry : scanner) {
    Key k = entry.getKey();
    String rowCol = k.getRow() + ":" + k.getColumnFamily() + ":" + k.getColumnQualifier() + ":"
        + String.format("%x", k.getTimestamp() & ColumnConstants.PREFIX_MASK);
    if (rowCols.contains(rowCol)) {
      System.err.println("DEBUG oldestTs : " + oldestTS + " recentTS : " + recentTS);
      Iterables.transform(scanner, e -> "DEBUG " + FluoFormatter.toString(e))
          .forEach(System.err::println);
    }
    Assert.assertFalse("Duplicate row col " + rowCol, rowCols.contains(rowCol));
    rowCols.add(rowCol);
  }
}
 

protected Text getTerm(Key key) {
  return key.getColumnFamily();
}
 

protected void findNextDocument() {
    topKey = null;
    
    try {
        Text cf = new Text();
        
        /*
         * Given that we are already at a document key, this method will continue to advance the underlying source until it is either exhausted (hasTop()
         * returns false), the returned key is not in the totalRange, and the current top key shares the same row and column family as the source's next
         * key.
         */
        while (topKey == null && source.hasTop()) {
            Key k = source.getTopKey();
            if (log.isTraceEnabled())
                log.trace("Sought to " + k);
            k.getColumnFamily(cf);
            
            if (!isEventKey(k)) {
                if (cf.find("fi\0") == 0) {
                    if (log.isDebugEnabled()) {
                        log.debug("Seeking over 'fi')");
                    }
                    // Try to do an optimized jump over the field index
                    cf.set("fi\1");
                    source.seek(new Range(new Key(source.getTopKey().getRow(), cf), false, totalRange.getEndKey(), totalRange.isEndKeyInclusive()),
                                    columnFamilies, inclusive);
                } else if (cf.getLength() == 1 && cf.charAt(0) == 'd') {
                    if (log.isDebugEnabled()) {
                        log.debug("Seeking over 'd'");
                    }
                    // Try to do an optimized jump over the raw documents
                    cf.set("d\0");
                    source.seek(new Range(new Key(source.getTopKey().getRow(), cf), false, totalRange.getEndKey(), totalRange.isEndKeyInclusive()),
                                    columnFamilies, inclusive);
                } else if (cf.getLength() == 2 && cf.charAt(0) == 't' && cf.charAt(1) == 'f') {
                    if (log.isDebugEnabled()) {
                        log.debug("Seeking over 'tf'");
                    }
                    // Try to do an optimized jump over the term frequencies
                    cf.set("tf\0");
                    source.seek(new Range(new Key(source.getTopKey().getRow(), cf), false, totalRange.getEndKey(), totalRange.isEndKeyInclusive()),
                                    columnFamilies, inclusive);
                } else {
                    if (log.isDebugEnabled()) {
                        log.debug("Next()'ing over the current key");
                    }
                    source.next();
                }
            } else {
                if (dataTypeFilter.apply(source.getTopKey())) {
                    this.topKey = source.getTopKey();
                } else {
                    Range nextCF = new Range(nextStartKey(source.getTopKey()), true, totalRange.getEndKey(), totalRange.isEndKeyInclusive());
                    source.seek(nextCF, columnFamilies, inclusive);
                }
            }
        }
    } catch (IOException e) {
        throw new RuntimeException("Could not seek in findNextDocument", e);
    }
}
 

@Override
protected Key getSeekStartKey(Key current, ByteSequence pointer) {
    // CQ = dataType\0UID\0Normalized field value\0Field name
    // seek to the next documents TF
    return new Key(current.getRow(), current.getColumnFamily(), new Text(pointer + Constants.NULL_BYTE_STRING + Constants.MAX_UNICODE_STRING));
}
 

@Override
public Key getStartKey(Key from) {
    return new Key(from.getRow(), from.getColumnFamily());
}
 

@Override
public Key getStartKey(Key from) {
    return new Key(from.getRow(), from.getColumnFamily());
}
 

protected Text getTerm(final Key key) {
	return key.getColumnFamily();
}
 

@Override
public Key getStartKey(Key from) {
    // don't delegate to the filter, we need an implementation here to override subclasses.
    return new Key(from.getRow(), from.getColumnFamily());
}
 

/**
 * Override this for your specific implementation.
 * 
 * In this specific implementation our row key will be the term, the column family will be the field name, and the column family will be the shard,so we
 * should have the following as our last key
 * 
 * bar FOO:20130101_0
 * 
 * so we should append a null so that we we don't skip shards. similarly, an assumption is made of the key structure within this class.
 * 
 * @param lastKey
 * @param previousRange
 */
public Range buildNextRange(final Key lastKey, final Range previousRange) {
    
    /**
     * This will re-seek the next column family when performing any field expansion.
     */
    Range r = new Range(new Key(lastKey.getRow(), new Text(lastKey.getColumnFamily() + "\u0000\uffff")), true, previousRange.getEndKey(),
                    previousRange.isEndKeyInclusive());
    if (log.isTraceEnabled())
        log.trace(r);
    return r;
    
}
 

/**
 * Seek starting from the end of the current field
 * 
 * @param current
 *            the current key
 * @param fieldName
 *            the field name to be seeked
 * @param endKey
 *            the current seek end key
 * @param endKeyInclusive
 *            the current seek end key inclusive flag
 * @return a new range that begins at the end of the current field
 */
private Range getFieldSeek(Key current, String fieldName, Key endKey, boolean endKeyInclusive) {
    Key startKey = new Key(current.getRow(), current.getColumnFamily(), new Text(fieldName + "\u0001"));
    return new Range(startKey, true, endKey, endKeyInclusive);
}
 

/**
 * Returns the given key's dataLocation
 *
 * @param key
 * @return The given key's dataLocation
 */
protected Text getDataLocation(Key key) {
    return key.getColumnFamily();
}