blob: dddf95837d0dc503a191abb2ba659e1ba22c8ea7 [file] [log] [blame]
<!DOCTYPE HTML>
<html lang="en">
<head>
<!-- Generated by javadoc (17) -->
<title>Source code</title>
<meta name="viewport" content="width=device-width, initial-scale=1">
<meta name="description" content="source: package: org.apache.hadoop.hbase.mapreduce, class: PutCombiner">
<meta name="generator" content="javadoc/SourceToHTMLConverter">
<link rel="stylesheet" type="text/css" href="../../../../../../stylesheet.css" title="Style">
</head>
<body class="source-page">
<main role="main">
<div class="source-container">
<pre><span class="source-line-no">001</span><span id="line-1">/*</span>
<span class="source-line-no">002</span><span id="line-2"> * Licensed to the Apache Software Foundation (ASF) under one</span>
<span class="source-line-no">003</span><span id="line-3"> * or more contributor license agreements. See the NOTICE file</span>
<span class="source-line-no">004</span><span id="line-4"> * distributed with this work for additional information</span>
<span class="source-line-no">005</span><span id="line-5"> * regarding copyright ownership. The ASF licenses this file</span>
<span class="source-line-no">006</span><span id="line-6"> * to you under the Apache License, Version 2.0 (the</span>
<span class="source-line-no">007</span><span id="line-7"> * "License"); you may not use this file except in compliance</span>
<span class="source-line-no">008</span><span id="line-8"> * with the License. You may obtain a copy of the License at</span>
<span class="source-line-no">009</span><span id="line-9"> *</span>
<span class="source-line-no">010</span><span id="line-10"> * http://www.apache.org/licenses/LICENSE-2.0</span>
<span class="source-line-no">011</span><span id="line-11"> *</span>
<span class="source-line-no">012</span><span id="line-12"> * Unless required by applicable law or agreed to in writing, software</span>
<span class="source-line-no">013</span><span id="line-13"> * distributed under the License is distributed on an "AS IS" BASIS,</span>
<span class="source-line-no">014</span><span id="line-14"> * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.</span>
<span class="source-line-no">015</span><span id="line-15"> * See the License for the specific language governing permissions and</span>
<span class="source-line-no">016</span><span id="line-16"> * limitations under the License.</span>
<span class="source-line-no">017</span><span id="line-17"> */</span>
<span class="source-line-no">018</span><span id="line-18">package org.apache.hadoop.hbase.mapreduce;</span>
<span class="source-line-no">019</span><span id="line-19"></span>
<span class="source-line-no">020</span><span id="line-20">import java.io.IOException;</span>
<span class="source-line-no">021</span><span id="line-21">import java.util.List;</span>
<span class="source-line-no">022</span><span id="line-22">import java.util.Map;</span>
<span class="source-line-no">023</span><span id="line-23">import java.util.Map.Entry;</span>
<span class="source-line-no">024</span><span id="line-24">import org.apache.hadoop.hbase.ExtendedCell;</span>
<span class="source-line-no">025</span><span id="line-25">import org.apache.hadoop.hbase.client.ClientInternalHelper;</span>
<span class="source-line-no">026</span><span id="line-26">import org.apache.hadoop.hbase.client.Put;</span>
<span class="source-line-no">027</span><span id="line-27">import org.apache.hadoop.mapreduce.Reducer;</span>
<span class="source-line-no">028</span><span id="line-28">import org.apache.yetus.audience.InterfaceAudience;</span>
<span class="source-line-no">029</span><span id="line-29">import org.slf4j.Logger;</span>
<span class="source-line-no">030</span><span id="line-30">import org.slf4j.LoggerFactory;</span>
<span class="source-line-no">031</span><span id="line-31"></span>
<span class="source-line-no">032</span><span id="line-32">/**</span>
<span class="source-line-no">033</span><span id="line-33"> * Combine Puts. Merges Put instances grouped by &lt;code&gt;K&lt;/code&gt; into a single instance.</span>
<span class="source-line-no">034</span><span id="line-34"> * @see TableMapReduceUtil</span>
<span class="source-line-no">035</span><span id="line-35"> */</span>
<span class="source-line-no">036</span><span id="line-36">@InterfaceAudience.Public</span>
<span class="source-line-no">037</span><span id="line-37">public class PutCombiner&lt;K&gt; extends Reducer&lt;K, Put, K, Put&gt; {</span>
<span class="source-line-no">038</span><span id="line-38"> private static final Logger LOG = LoggerFactory.getLogger(PutCombiner.class);</span>
<span class="source-line-no">039</span><span id="line-39"></span>
<span class="source-line-no">040</span><span id="line-40"> @Override</span>
<span class="source-line-no">041</span><span id="line-41"> protected void reduce(K row, Iterable&lt;Put&gt; vals, Context context)</span>
<span class="source-line-no">042</span><span id="line-42"> throws IOException, InterruptedException {</span>
<span class="source-line-no">043</span><span id="line-43"> // Using HeapSize to create an upper bound on the memory size of</span>
<span class="source-line-no">044</span><span id="line-44"> // the puts and flush some portion of the content while looping. This</span>
<span class="source-line-no">045</span><span id="line-45"> // flush could result in multiple Puts for a single rowkey. That is</span>
<span class="source-line-no">046</span><span id="line-46"> // acceptable because Combiner is run as an optimization and it's not</span>
<span class="source-line-no">047</span><span id="line-47"> // critical that all Puts are grouped perfectly.</span>
<span class="source-line-no">048</span><span id="line-48"> long threshold =</span>
<span class="source-line-no">049</span><span id="line-49"> context.getConfiguration().getLong("putcombiner.row.threshold", 1L * (1 &lt;&lt; 30));</span>
<span class="source-line-no">050</span><span id="line-50"> int cnt = 0;</span>
<span class="source-line-no">051</span><span id="line-51"> long curSize = 0;</span>
<span class="source-line-no">052</span><span id="line-52"> Put combinedPut = null;</span>
<span class="source-line-no">053</span><span id="line-53"> Map&lt;byte[], List&lt;ExtendedCell&gt;&gt; combinedFamilyMap = null;</span>
<span class="source-line-no">054</span><span id="line-54"> for (Put p : vals) {</span>
<span class="source-line-no">055</span><span id="line-55"> cnt++;</span>
<span class="source-line-no">056</span><span id="line-56"> if (combinedPut == null) {</span>
<span class="source-line-no">057</span><span id="line-57"> combinedPut = p;</span>
<span class="source-line-no">058</span><span id="line-58"> combinedFamilyMap = ClientInternalHelper.getExtendedFamilyCellMap(combinedPut);</span>
<span class="source-line-no">059</span><span id="line-59"> } else {</span>
<span class="source-line-no">060</span><span id="line-60"> for (Entry&lt;byte[], List&lt;ExtendedCell&gt;&gt; entry : ClientInternalHelper</span>
<span class="source-line-no">061</span><span id="line-61"> .getExtendedFamilyCellMap(p).entrySet()) {</span>
<span class="source-line-no">062</span><span id="line-62"> List&lt;ExtendedCell&gt; existCells = combinedFamilyMap.get(entry.getKey());</span>
<span class="source-line-no">063</span><span id="line-63"> if (existCells == null) {</span>
<span class="source-line-no">064</span><span id="line-64"> // no cells for this family yet, just put it</span>
<span class="source-line-no">065</span><span id="line-65"> combinedFamilyMap.put(entry.getKey(), entry.getValue());</span>
<span class="source-line-no">066</span><span id="line-66"> // do not forget to calculate the size</span>
<span class="source-line-no">067</span><span id="line-67"> for (ExtendedCell cell : entry.getValue()) {</span>
<span class="source-line-no">068</span><span id="line-68"> curSize += cell.heapSize();</span>
<span class="source-line-no">069</span><span id="line-69"> }</span>
<span class="source-line-no">070</span><span id="line-70"> } else {</span>
<span class="source-line-no">071</span><span id="line-71"> // otherwise just add the cells to the existent list for this family</span>
<span class="source-line-no">072</span><span id="line-72"> for (ExtendedCell cell : entry.getValue()) {</span>
<span class="source-line-no">073</span><span id="line-73"> existCells.add(cell);</span>
<span class="source-line-no">074</span><span id="line-74"> curSize += cell.heapSize();</span>
<span class="source-line-no">075</span><span id="line-75"> }</span>
<span class="source-line-no">076</span><span id="line-76"> }</span>
<span class="source-line-no">077</span><span id="line-77"> }</span>
<span class="source-line-no">078</span><span id="line-78"> if (cnt % 10 == 0) {</span>
<span class="source-line-no">079</span><span id="line-79"> context.setStatus("Combine " + cnt);</span>
<span class="source-line-no">080</span><span id="line-80"> }</span>
<span class="source-line-no">081</span><span id="line-81"> if (curSize &gt; threshold) {</span>
<span class="source-line-no">082</span><span id="line-82"> if (LOG.isDebugEnabled()) {</span>
<span class="source-line-no">083</span><span id="line-83"> LOG.debug(String.format("Combined %d Put(s) into %d.", cnt, 1));</span>
<span class="source-line-no">084</span><span id="line-84"> }</span>
<span class="source-line-no">085</span><span id="line-85"> context.write(row, combinedPut);</span>
<span class="source-line-no">086</span><span id="line-86"> combinedPut = null;</span>
<span class="source-line-no">087</span><span id="line-87"> curSize = 0;</span>
<span class="source-line-no">088</span><span id="line-88"> cnt = 0;</span>
<span class="source-line-no">089</span><span id="line-89"> }</span>
<span class="source-line-no">090</span><span id="line-90"> }</span>
<span class="source-line-no">091</span><span id="line-91"> }</span>
<span class="source-line-no">092</span><span id="line-92"> if (combinedPut != null) {</span>
<span class="source-line-no">093</span><span id="line-93"> if (LOG.isDebugEnabled()) {</span>
<span class="source-line-no">094</span><span id="line-94"> LOG.debug(String.format("Combined %d Put(s) into %d.", cnt, 1));</span>
<span class="source-line-no">095</span><span id="line-95"> }</span>
<span class="source-line-no">096</span><span id="line-96"> context.write(row, combinedPut);</span>
<span class="source-line-no">097</span><span id="line-97"> }</span>
<span class="source-line-no">098</span><span id="line-98"> }</span>
<span class="source-line-no">099</span><span id="line-99">}</span>
</pre>
</div>
</main>
</body>
</html>