001/*
002 * Licensed to the Apache Software Foundation (ASF) under one
003 * or more contributor license agreements.  See the NOTICE file
004 * distributed with this work for additional information
005 * regarding copyright ownership.  The ASF licenses this file
006 * to you under the Apache License, Version 2.0 (the
007 * "License"); you may not use this file except in compliance
008 * with the License.  You may obtain a copy of the License at
009 *
010 *     http://www.apache.org/licenses/LICENSE-2.0
011 *
012 * Unless required by applicable law or agreed to in writing, software
013 * distributed under the License is distributed on an "AS IS" BASIS,
014 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
015 * See the License for the specific language governing permissions and
016 * limitations under the License.
017 */
018package org.apache.hadoop.hbase.mapreduce;
019
020import java.io.IOException;
021import org.apache.hadoop.conf.Configuration;
022import org.apache.hadoop.conf.Configured;
023import org.apache.hadoop.fs.Path;
024import org.apache.hadoop.hbase.Cell;
025import org.apache.hadoop.hbase.CellUtil;
026import org.apache.hadoop.hbase.CompareOperator;
027import org.apache.hadoop.hbase.HBaseConfiguration;
028import org.apache.hadoop.hbase.HConstants;
029import org.apache.hadoop.hbase.client.Result;
030import org.apache.hadoop.hbase.client.Scan;
031import org.apache.hadoop.hbase.filter.Filter;
032import org.apache.hadoop.hbase.filter.PrefixFilter;
033import org.apache.hadoop.hbase.filter.RegexStringComparator;
034import org.apache.hadoop.hbase.filter.RowFilter;
035import org.apache.hadoop.hbase.io.ImmutableBytesWritable;
036import org.apache.hadoop.hbase.util.Bytes;
037import org.apache.hadoop.io.LongWritable;
038import org.apache.hadoop.io.Text;
039import org.apache.hadoop.mapreduce.Job;
040import org.apache.hadoop.mapreduce.Reducer;
041import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat;
042import org.apache.hadoop.mapreduce.lib.output.TextOutputFormat;
043import org.apache.hadoop.util.Tool;
044import org.apache.hadoop.util.ToolRunner;
045import org.apache.yetus.audience.InterfaceAudience;
046import org.slf4j.Logger;
047import org.slf4j.LoggerFactory;
048
049import org.apache.hbase.thirdparty.com.google.common.base.Preconditions;
050
051/**
052 * A job with a a map and reduce phase to count cells in a table. The counter lists the following
053 * stats for a given table:
054 *
055 * <pre>
056 * 1. Total number of rows in the table
057 * 2. Total number of CFs across all rows
058 * 3. Total qualifiers across all rows
059 * 4. Total occurrence of each CF
060 * 5. Total occurrence of each qualifier
061 * 6. Total number of versions of each qualifier.
062 * 7. Total size of serialized cells of each CF.
063 * 8. Total size of serialized cells of each qualifier.
064 * 9. Total size of serialized cells across all rows.
065 * </pre>
066 *
067 * The cellcounter can take optional parameters to use a user supplied row/family/qualifier string
068 * to use in the report and second a regex based or prefix based row filter to restrict the count
069 * operation to a limited subset of rows from the table or a start time and/or end time to limit the
070 * count to a time range.
071 */
072@InterfaceAudience.Public
073public class CellCounter extends Configured implements Tool {
074  private static final Logger LOG = LoggerFactory.getLogger(CellCounter.class.getName());
075
076  /**
077   * Name of this 'program'.
078   */
079  static final String NAME = "CellCounter";
080
081  private final static String JOB_NAME_CONF_KEY = "mapreduce.job.name";
082
083  /**
084   * Mapper that runs the count.
085   */
086  static class CellCounterMapper extends TableMapper<Text, LongWritable> {
087    /**
088     * Counter enumeration to count the actual rows.
089     */
090    public static enum Counters {
091      ROWS,
092      CELLS,
093      SIZE
094    }
095
096    private Configuration conf;
097    private String separator;
098
099    // state of current row, family, column needs to persist across map() invocations
100    // in order to properly handle scanner batching, where a single qualifier may have too
101    // many versions for a single map() call
102    private byte[] lastRow;
103    private String currentRowKey;
104    byte[] currentFamily = null;
105    String currentFamilyName = null;
106    byte[] currentQualifier = null;
107    // family + qualifier
108    String currentQualifierName = null;
109    // rowkey + family + qualifier
110    String currentRowQualifierName = null;
111
112    @Override
113    protected void setup(Context context) throws IOException, InterruptedException {
114      conf = context.getConfiguration();
115      separator = conf.get("ReportSeparator", ":");
116    }
117
118    /**
119     * Maps the data.
120     * @param row     The current table row key.
121     * @param values  The columns.
122     * @param context The current context.
123     * @throws IOException When something is broken with the data.
124     */
125
126    @Override
127    @edu.umd.cs.findbugs.annotations.SuppressWarnings(value = "NP_NULL_ON_SOME_PATH",
128        justification = "Findbugs is blind to the Precondition null check")
129    public void map(ImmutableBytesWritable row, Result values, Context context) throws IOException {
130      Preconditions.checkState(values != null, "values passed to the map is null");
131
132      try {
133        byte[] currentRow = values.getRow();
134        if (lastRow == null || !Bytes.equals(lastRow, currentRow)) {
135          lastRow = currentRow;
136          currentRowKey = Bytes.toStringBinary(currentRow);
137          currentFamily = null;
138          currentQualifier = null;
139          context.getCounter(Counters.ROWS).increment(1);
140          context.write(new Text("Total ROWS"), new LongWritable(1));
141        }
142        if (!values.isEmpty()) {
143          int cellCount = 0;
144          for (Cell value : values.listCells()) {
145            cellCount++;
146            long size = value.getSerializedSize();
147            if (currentFamily == null || !CellUtil.matchingFamily(value, currentFamily)) {
148              currentFamily = CellUtil.cloneFamily(value);
149              currentFamilyName = Bytes.toStringBinary(currentFamily);
150              currentQualifier = null;
151              context.getCounter("CF", currentFamilyName).increment(1);
152              if (1 == context.getCounter("CF", currentFamilyName).getValue()) {
153                context.write(new Text("Total Families Across all Rows"), new LongWritable(1));
154                context.write(new Text(currentFamily), new LongWritable(1));
155              }
156              context.getCounter(Counters.SIZE).increment(size);
157              context.write(new Text("Total SIZE"), new LongWritable(size));
158              context.getCounter("CF", currentFamilyName + "_Size").increment(size);
159              context.write(new Text(currentFamilyName + "_Size"), new LongWritable(size));
160            }
161            if (currentQualifier == null || !CellUtil.matchingQualifier(value, currentQualifier)) {
162              currentQualifier = CellUtil.cloneQualifier(value);
163              currentQualifierName =
164                currentFamilyName + separator + Bytes.toStringBinary(currentQualifier);
165              currentRowQualifierName = currentRowKey + separator + currentQualifierName;
166
167              context.write(new Text("Total Qualifiers across all Rows"), new LongWritable(1));
168              context.write(new Text(currentQualifierName), new LongWritable(1));
169              context.getCounter("Q", currentQualifierName + "_Size").increment(size);
170              context.write(new Text(currentQualifierName + "_Size"), new LongWritable(size));
171            }
172            // Increment versions
173            context.write(new Text(currentRowQualifierName + "_Versions"), new LongWritable(1));
174          }
175          context.getCounter(Counters.CELLS).increment(cellCount);
176        }
177      } catch (InterruptedException e) {
178        LOG.error("Interrupted while writing cellCount", e);
179        Thread.currentThread().interrupt();
180      }
181    }
182  }
183
184  static class LongSumReducer<Key> extends Reducer<Key, LongWritable, Key, LongWritable> {
185
186    private LongWritable result = new LongWritable();
187
188    public void reduce(Key key, Iterable<LongWritable> values, Context context)
189      throws IOException, InterruptedException {
190      long sum = 0;
191      for (LongWritable val : values) {
192        sum += val.get();
193      }
194      result.set(sum);
195      context.write(key, result);
196    }
197
198  }
199
200  /**
201   * Sets up the actual job.
202   * @param conf The current configuration.
203   * @param args The command line parameters.
204   * @return The newly created job.
205   * @throws IOException When setting up the job fails.
206   */
207  public static Job createSubmittableJob(Configuration conf, String[] args) throws IOException {
208    String tableName = args[0];
209    Path outputDir = new Path(args[1]);
210    String reportSeparatorString = (args.length > 2) ? args[2] : ":";
211    conf.set("ReportSeparator", reportSeparatorString);
212    Job job = Job.getInstance(conf, conf.get(JOB_NAME_CONF_KEY, NAME + "_" + tableName));
213    job.setJarByClass(CellCounter.class);
214    Scan scan = getConfiguredScanForJob(conf, args);
215    TableMapReduceUtil.initTableMapperJob(tableName, scan, CellCounterMapper.class,
216      ImmutableBytesWritable.class, Result.class, job);
217    job.setMapOutputKeyClass(Text.class);
218    job.setMapOutputValueClass(LongWritable.class);
219    job.setOutputFormatClass(TextOutputFormat.class);
220    job.setOutputKeyClass(Text.class);
221    job.setOutputValueClass(LongWritable.class);
222    FileOutputFormat.setOutputPath(job, outputDir);
223    job.setReducerClass(LongSumReducer.class);
224    job.setCombinerClass(LongSumReducer.class);
225    return job;
226  }
227
228  private static Scan getConfiguredScanForJob(Configuration conf, String[] args)
229    throws IOException {
230    // create scan with any properties set from TableInputFormat
231    Scan s = TableInputFormat.createScanFromConfiguration(conf);
232    // Set Scan Versions
233    if (conf.get(TableInputFormat.SCAN_MAXVERSIONS) == null) {
234      // default to all versions unless explicitly set
235      s.readVersions(Integer.MAX_VALUE);
236    }
237    s.setCacheBlocks(false);
238    // Set RowFilter or Prefix Filter if applicable.
239    Filter rowFilter = getRowFilter(args);
240    if (rowFilter != null) {
241      LOG.info("Setting Row Filter for counter.");
242      s.setFilter(rowFilter);
243    }
244    // Set TimeRange if defined
245    long timeRange[] = getTimeRange(args);
246    if (timeRange != null) {
247      LOG.info("Setting TimeRange for counter.");
248      s.setTimeRange(timeRange[0], timeRange[1]);
249    }
250    return s;
251  }
252
253  private static Filter getRowFilter(String[] args) {
254    Filter rowFilter = null;
255    String filterCriteria = (args.length > 3) ? args[3] : null;
256    if (filterCriteria == null) return null;
257    if (filterCriteria.startsWith("^")) {
258      String regexPattern = filterCriteria.substring(1, filterCriteria.length());
259      rowFilter = new RowFilter(CompareOperator.EQUAL, new RegexStringComparator(regexPattern));
260    } else {
261      rowFilter = new PrefixFilter(Bytes.toBytesBinary(filterCriteria));
262    }
263    return rowFilter;
264  }
265
266  private static long[] getTimeRange(String[] args) throws IOException {
267    final String startTimeArgKey = "--starttime=";
268    final String endTimeArgKey = "--endtime=";
269    long startTime = 0L;
270    long endTime = 0L;
271
272    for (int i = 1; i < args.length; i++) {
273      System.out.println("i:" + i + "arg[i]" + args[i]);
274      if (args[i].startsWith(startTimeArgKey)) {
275        startTime = Long.parseLong(args[i].substring(startTimeArgKey.length()));
276      }
277      if (args[i].startsWith(endTimeArgKey)) {
278        endTime = Long.parseLong(args[i].substring(endTimeArgKey.length()));
279      }
280    }
281
282    if (startTime == 0 && endTime == 0) return null;
283
284    endTime = endTime == 0 ? HConstants.LATEST_TIMESTAMP : endTime;
285    return new long[] { startTime, endTime };
286  }
287
288  @Override
289  public int run(String[] args) throws Exception {
290    if (args.length < 2) {
291      printUsage(args.length);
292      return -1;
293    }
294    Job job = createSubmittableJob(getConf(), args);
295    return (job.waitForCompletion(true) ? 0 : 1);
296  }
297
298  private void printUsage(int parameterCount) {
299    System.err.println("ERROR: Wrong number of parameters: " + parameterCount);
300    System.err.println("Usage: hbase cellcounter <tablename> <outputDir> [reportSeparator] "
301      + "[^[regex pattern] or [Prefix]] [--starttime=<starttime> --endtime=<endtime>]");
302    System.err.println("  Note: -D properties will be applied to the conf used.");
303    System.err.println("  Additionally, all of the SCAN properties from TableInputFormat can be "
304      + "specified to get fine grained control on what is counted.");
305    System.err.println("   -D" + TableInputFormat.SCAN_ROW_START + "=<rowkey>");
306    System.err.println("   -D" + TableInputFormat.SCAN_ROW_STOP + "=<rowkey>");
307    System.err.println("   -D" + TableInputFormat.SCAN_COLUMNS + "=\"<col1> <col2>...\"");
308    System.err.println("   -D" + TableInputFormat.SCAN_COLUMN_FAMILY + "=<family1>,<family2>, ...");
309    System.err.println("   -D" + TableInputFormat.SCAN_TIMESTAMP + "=<timestamp>");
310    System.err.println("   -D" + TableInputFormat.SCAN_TIMERANGE_START + "=<timestamp>");
311    System.err.println("   -D" + TableInputFormat.SCAN_TIMERANGE_END + "=<timestamp>");
312    System.err.println("   -D" + TableInputFormat.SCAN_MAXVERSIONS + "=<count>");
313    System.err.println("   -D" + TableInputFormat.SCAN_CACHEDROWS + "=<count>");
314    System.err.println("   -D" + TableInputFormat.SCAN_BATCHSIZE + "=<count>");
315    System.err.println(" <reportSeparator> parameter can be used to override the default report "
316      + "separator string : used to separate the rowId/column family name and qualifier name.");
317    System.err.println(" [^[regex pattern] or [Prefix] parameter can be used to limit the cell "
318      + "counter count operation to a limited subset of rows from the table based on regex or "
319      + "prefix pattern.");
320  }
321
322  /**
323   * Main entry point.
324   * @param args The command line parameters.
325   * @throws Exception When running the job fails.
326   */
327  public static void main(String[] args) throws Exception {
328    int errCode = ToolRunner.run(HBaseConfiguration.create(), new CellCounter(), args);
329    System.exit(errCode);
330  }
331
332}