blob: db30d203bdddff93fd05cd5e1b9b3fb9eff60ea0 [file] [log] [blame]
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd">
<html lang="en">
<head>
<title>Source code</title>
<link rel="stylesheet" type="text/css" href="../../../../../../stylesheet.css" title="Style">
</head>
<body>
<div class="sourceContainer">
<pre><span class="sourceLineNo">001</span>/**<a name="line.1"></a>
<span class="sourceLineNo">002</span> *<a name="line.2"></a>
<span class="sourceLineNo">003</span> * Licensed to the Apache Software Foundation (ASF) under one<a name="line.3"></a>
<span class="sourceLineNo">004</span> * or more contributor license agreements. See the NOTICE file<a name="line.4"></a>
<span class="sourceLineNo">005</span> * distributed with this work for additional information<a name="line.5"></a>
<span class="sourceLineNo">006</span> * regarding copyright ownership. The ASF licenses this file<a name="line.6"></a>
<span class="sourceLineNo">007</span> * to you under the Apache License, Version 2.0 (the<a name="line.7"></a>
<span class="sourceLineNo">008</span> * "License"); you may not use this file except in compliance<a name="line.8"></a>
<span class="sourceLineNo">009</span> * with the License. You may obtain a copy of the License at<a name="line.9"></a>
<span class="sourceLineNo">010</span> *<a name="line.10"></a>
<span class="sourceLineNo">011</span> * http://www.apache.org/licenses/LICENSE-2.0<a name="line.11"></a>
<span class="sourceLineNo">012</span> *<a name="line.12"></a>
<span class="sourceLineNo">013</span> * Unless required by applicable law or agreed to in writing, software<a name="line.13"></a>
<span class="sourceLineNo">014</span> * distributed under the License is distributed on an "AS IS" BASIS,<a name="line.14"></a>
<span class="sourceLineNo">015</span> * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.<a name="line.15"></a>
<span class="sourceLineNo">016</span> * See the License for the specific language governing permissions and<a name="line.16"></a>
<span class="sourceLineNo">017</span> * limitations under the License.<a name="line.17"></a>
<span class="sourceLineNo">018</span> */<a name="line.18"></a>
<span class="sourceLineNo">019</span>package org.apache.hadoop.hbase.mapreduce;<a name="line.19"></a>
<span class="sourceLineNo">020</span><a name="line.20"></a>
<span class="sourceLineNo">021</span>import java.io.IOException;<a name="line.21"></a>
<span class="sourceLineNo">022</span>import java.util.List;<a name="line.22"></a>
<span class="sourceLineNo">023</span>import java.util.Map.Entry;<a name="line.23"></a>
<span class="sourceLineNo">024</span>import java.util.Map;<a name="line.24"></a>
<span class="sourceLineNo">025</span><a name="line.25"></a>
<span class="sourceLineNo">026</span>import org.apache.yetus.audience.InterfaceAudience;<a name="line.26"></a>
<span class="sourceLineNo">027</span>import org.slf4j.Logger;<a name="line.27"></a>
<span class="sourceLineNo">028</span>import org.slf4j.LoggerFactory;<a name="line.28"></a>
<span class="sourceLineNo">029</span>import org.apache.hadoop.hbase.Cell;<a name="line.29"></a>
<span class="sourceLineNo">030</span>import org.apache.hadoop.hbase.KeyValue;<a name="line.30"></a>
<span class="sourceLineNo">031</span>import org.apache.hadoop.hbase.KeyValueUtil;<a name="line.31"></a>
<span class="sourceLineNo">032</span>import org.apache.hadoop.hbase.client.Put;<a name="line.32"></a>
<span class="sourceLineNo">033</span>import org.apache.hadoop.mapreduce.Reducer;<a name="line.33"></a>
<span class="sourceLineNo">034</span><a name="line.34"></a>
<span class="sourceLineNo">035</span>/**<a name="line.35"></a>
<span class="sourceLineNo">036</span> * Combine Puts. Merges Put instances grouped by &lt;code&gt;K&lt;/code&gt; into a single<a name="line.36"></a>
<span class="sourceLineNo">037</span> * instance.<a name="line.37"></a>
<span class="sourceLineNo">038</span> * @see TableMapReduceUtil<a name="line.38"></a>
<span class="sourceLineNo">039</span> */<a name="line.39"></a>
<span class="sourceLineNo">040</span>@InterfaceAudience.Public<a name="line.40"></a>
<span class="sourceLineNo">041</span>public class PutCombiner&lt;K&gt; extends Reducer&lt;K, Put, K, Put&gt; {<a name="line.41"></a>
<span class="sourceLineNo">042</span> private static final Logger LOG = LoggerFactory.getLogger(PutCombiner.class);<a name="line.42"></a>
<span class="sourceLineNo">043</span><a name="line.43"></a>
<span class="sourceLineNo">044</span> @Override<a name="line.44"></a>
<span class="sourceLineNo">045</span> protected void reduce(K row, Iterable&lt;Put&gt; vals, Context context)<a name="line.45"></a>
<span class="sourceLineNo">046</span> throws IOException, InterruptedException {<a name="line.46"></a>
<span class="sourceLineNo">047</span> // Using HeapSize to create an upper bound on the memory size of<a name="line.47"></a>
<span class="sourceLineNo">048</span> // the puts and flush some portion of the content while looping. This<a name="line.48"></a>
<span class="sourceLineNo">049</span> // flush could result in multiple Puts for a single rowkey. That is<a name="line.49"></a>
<span class="sourceLineNo">050</span> // acceptable because Combiner is run as an optimization and it's not<a name="line.50"></a>
<span class="sourceLineNo">051</span> // critical that all Puts are grouped perfectly.<a name="line.51"></a>
<span class="sourceLineNo">052</span> long threshold = context.getConfiguration().getLong(<a name="line.52"></a>
<span class="sourceLineNo">053</span> "putcombiner.row.threshold", 1L * (1&lt;&lt;30));<a name="line.53"></a>
<span class="sourceLineNo">054</span> int cnt = 0;<a name="line.54"></a>
<span class="sourceLineNo">055</span> long curSize = 0;<a name="line.55"></a>
<span class="sourceLineNo">056</span> Put put = null;<a name="line.56"></a>
<span class="sourceLineNo">057</span> Map&lt;byte[], List&lt;Cell&gt;&gt; familyMap = null;<a name="line.57"></a>
<span class="sourceLineNo">058</span> for (Put p : vals) {<a name="line.58"></a>
<span class="sourceLineNo">059</span> cnt++;<a name="line.59"></a>
<span class="sourceLineNo">060</span> if (put == null) {<a name="line.60"></a>
<span class="sourceLineNo">061</span> put = p;<a name="line.61"></a>
<span class="sourceLineNo">062</span> familyMap = put.getFamilyCellMap();<a name="line.62"></a>
<span class="sourceLineNo">063</span> } else {<a name="line.63"></a>
<span class="sourceLineNo">064</span> for (Entry&lt;byte[], List&lt;Cell&gt;&gt; entry : p.getFamilyCellMap()<a name="line.64"></a>
<span class="sourceLineNo">065</span> .entrySet()) {<a name="line.65"></a>
<span class="sourceLineNo">066</span> List&lt;Cell&gt; cells = familyMap.get(entry.getKey());<a name="line.66"></a>
<span class="sourceLineNo">067</span> List&lt;Cell&gt; kvs = (cells != null) ? (List&lt;Cell&gt;) cells : null;<a name="line.67"></a>
<span class="sourceLineNo">068</span> for (Cell cell : entry.getValue()) {<a name="line.68"></a>
<span class="sourceLineNo">069</span> KeyValue kv = KeyValueUtil.ensureKeyValue(cell);<a name="line.69"></a>
<span class="sourceLineNo">070</span> curSize += kv.heapSize();<a name="line.70"></a>
<span class="sourceLineNo">071</span> if (kvs != null) {<a name="line.71"></a>
<span class="sourceLineNo">072</span> kvs.add(kv);<a name="line.72"></a>
<span class="sourceLineNo">073</span> }<a name="line.73"></a>
<span class="sourceLineNo">074</span> }<a name="line.74"></a>
<span class="sourceLineNo">075</span> if (cells == null) {<a name="line.75"></a>
<span class="sourceLineNo">076</span> familyMap.put(entry.getKey(), entry.getValue());<a name="line.76"></a>
<span class="sourceLineNo">077</span> }<a name="line.77"></a>
<span class="sourceLineNo">078</span> }<a name="line.78"></a>
<span class="sourceLineNo">079</span> if (cnt % 10 == 0) context.setStatus("Combine " + cnt);<a name="line.79"></a>
<span class="sourceLineNo">080</span> if (curSize &gt; threshold) {<a name="line.80"></a>
<span class="sourceLineNo">081</span> if (LOG.isDebugEnabled()) {<a name="line.81"></a>
<span class="sourceLineNo">082</span> LOG.debug(String.format("Combined %d Put(s) into %d.", cnt, 1));<a name="line.82"></a>
<span class="sourceLineNo">083</span> }<a name="line.83"></a>
<span class="sourceLineNo">084</span> context.write(row, put);<a name="line.84"></a>
<span class="sourceLineNo">085</span> put = null;<a name="line.85"></a>
<span class="sourceLineNo">086</span> curSize = 0;<a name="line.86"></a>
<span class="sourceLineNo">087</span> cnt = 0;<a name="line.87"></a>
<span class="sourceLineNo">088</span> }<a name="line.88"></a>
<span class="sourceLineNo">089</span> }<a name="line.89"></a>
<span class="sourceLineNo">090</span> }<a name="line.90"></a>
<span class="sourceLineNo">091</span> if (put != null) {<a name="line.91"></a>
<span class="sourceLineNo">092</span> if (LOG.isDebugEnabled()) {<a name="line.92"></a>
<span class="sourceLineNo">093</span> LOG.debug(String.format("Combined %d Put(s) into %d.", cnt, 1));<a name="line.93"></a>
<span class="sourceLineNo">094</span> }<a name="line.94"></a>
<span class="sourceLineNo">095</span> context.write(row, put);<a name="line.95"></a>
<span class="sourceLineNo">096</span> }<a name="line.96"></a>
<span class="sourceLineNo">097</span> }<a name="line.97"></a>
<span class="sourceLineNo">098</span>}<a name="line.98"></a>
</pre>
</div>
</body>
</html>