| <!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 org.apache.yetus.audience.InterfaceAudience;<a name="line.22"></a> |
| <span class="sourceLineNo">023</span>import org.slf4j.Logger;<a name="line.23"></a> |
| <span class="sourceLineNo">024</span>import org.slf4j.LoggerFactory;<a name="line.24"></a> |
| <span class="sourceLineNo">025</span>import org.apache.hadoop.conf.Configurable;<a name="line.25"></a> |
| <span class="sourceLineNo">026</span>import org.apache.hadoop.conf.Configuration;<a name="line.26"></a> |
| <span class="sourceLineNo">027</span>import org.apache.hadoop.hbase.HBaseConfiguration;<a name="line.27"></a> |
| <span class="sourceLineNo">028</span>import org.apache.hadoop.hbase.TableName;<a name="line.28"></a> |
| <span class="sourceLineNo">029</span>import org.apache.hadoop.hbase.client.Connection;<a name="line.29"></a> |
| <span class="sourceLineNo">030</span>import org.apache.hadoop.hbase.client.ConnectionFactory;<a name="line.30"></a> |
| <span class="sourceLineNo">031</span>import org.apache.hadoop.hbase.client.RegionLocator;<a name="line.31"></a> |
| <span class="sourceLineNo">032</span>import org.apache.hadoop.hbase.io.ImmutableBytesWritable;<a name="line.32"></a> |
| <span class="sourceLineNo">033</span>import org.apache.hadoop.hbase.mapred.TableOutputFormat;<a name="line.33"></a> |
| <span class="sourceLineNo">034</span>import org.apache.hadoop.hbase.util.Bytes;<a name="line.34"></a> |
| <span class="sourceLineNo">035</span>import org.apache.hadoop.mapreduce.Partitioner;<a name="line.35"></a> |
| <span class="sourceLineNo">036</span><a name="line.36"></a> |
| <span class="sourceLineNo">037</span>/**<a name="line.37"></a> |
| <span class="sourceLineNo">038</span> * This is used to partition the output keys into groups of keys.<a name="line.38"></a> |
| <span class="sourceLineNo">039</span> * Keys are grouped according to the regions that currently exist<a name="line.39"></a> |
| <span class="sourceLineNo">040</span> * so that each reducer fills a single region so load is distributed.<a name="line.40"></a> |
| <span class="sourceLineNo">041</span> *<a name="line.41"></a> |
| <span class="sourceLineNo">042</span> * <p>This class is not suitable as partitioner creating hfiles<a name="line.42"></a> |
| <span class="sourceLineNo">043</span> * for incremental bulk loads as region spread will likely change between time of<a name="line.43"></a> |
| <span class="sourceLineNo">044</span> * hfile creation and load time. See {@link org.apache.hadoop.hbase.tool.BulkLoadHFiles}<a name="line.44"></a> |
| <span class="sourceLineNo">045</span> * and <a href="http://hbase.apache.org/book.html#arch.bulk.load">Bulk Load</a>.</p><a name="line.45"></a> |
| <span class="sourceLineNo">046</span> *<a name="line.46"></a> |
| <span class="sourceLineNo">047</span> * @param <KEY> The type of the key.<a name="line.47"></a> |
| <span class="sourceLineNo">048</span> * @param <VALUE> The type of the value.<a name="line.48"></a> |
| <span class="sourceLineNo">049</span> */<a name="line.49"></a> |
| <span class="sourceLineNo">050</span>@InterfaceAudience.Public<a name="line.50"></a> |
| <span class="sourceLineNo">051</span>public class HRegionPartitioner<KEY, VALUE><a name="line.51"></a> |
| <span class="sourceLineNo">052</span>extends Partitioner<ImmutableBytesWritable, VALUE><a name="line.52"></a> |
| <span class="sourceLineNo">053</span>implements Configurable {<a name="line.53"></a> |
| <span class="sourceLineNo">054</span><a name="line.54"></a> |
| <span class="sourceLineNo">055</span> private static final Logger LOG = LoggerFactory.getLogger(HRegionPartitioner.class);<a name="line.55"></a> |
| <span class="sourceLineNo">056</span> private Configuration conf = null;<a name="line.56"></a> |
| <span class="sourceLineNo">057</span> // Connection and locator are not cleaned up; they just die when partitioner is done.<a name="line.57"></a> |
| <span class="sourceLineNo">058</span> private Connection connection;<a name="line.58"></a> |
| <span class="sourceLineNo">059</span> private RegionLocator locator;<a name="line.59"></a> |
| <span class="sourceLineNo">060</span> private byte[][] startKeys;<a name="line.60"></a> |
| <span class="sourceLineNo">061</span><a name="line.61"></a> |
| <span class="sourceLineNo">062</span> /**<a name="line.62"></a> |
| <span class="sourceLineNo">063</span> * Gets the partition number for a given key (hence record) given the total<a name="line.63"></a> |
| <span class="sourceLineNo">064</span> * number of partitions i.e. number of reduce-tasks for the job.<a name="line.64"></a> |
| <span class="sourceLineNo">065</span> *<a name="line.65"></a> |
| <span class="sourceLineNo">066</span> * <p>Typically a hash function on a all or a subset of the key.</p><a name="line.66"></a> |
| <span class="sourceLineNo">067</span> *<a name="line.67"></a> |
| <span class="sourceLineNo">068</span> * @param key The key to be partitioned.<a name="line.68"></a> |
| <span class="sourceLineNo">069</span> * @param value The entry value.<a name="line.69"></a> |
| <span class="sourceLineNo">070</span> * @param numPartitions The total number of partitions.<a name="line.70"></a> |
| <span class="sourceLineNo">071</span> * @return The partition number for the <code>key</code>.<a name="line.71"></a> |
| <span class="sourceLineNo">072</span> * @see org.apache.hadoop.mapreduce.Partitioner#getPartition(<a name="line.72"></a> |
| <span class="sourceLineNo">073</span> * java.lang.Object, java.lang.Object, int)<a name="line.73"></a> |
| <span class="sourceLineNo">074</span> */<a name="line.74"></a> |
| <span class="sourceLineNo">075</span> @Override<a name="line.75"></a> |
| <span class="sourceLineNo">076</span> public int getPartition(ImmutableBytesWritable key,<a name="line.76"></a> |
| <span class="sourceLineNo">077</span> VALUE value, int numPartitions) {<a name="line.77"></a> |
| <span class="sourceLineNo">078</span> byte[] region = null;<a name="line.78"></a> |
| <span class="sourceLineNo">079</span> // Only one region return 0<a name="line.79"></a> |
| <span class="sourceLineNo">080</span> if (this.startKeys.length == 1){<a name="line.80"></a> |
| <span class="sourceLineNo">081</span> return 0;<a name="line.81"></a> |
| <span class="sourceLineNo">082</span> }<a name="line.82"></a> |
| <span class="sourceLineNo">083</span> try {<a name="line.83"></a> |
| <span class="sourceLineNo">084</span> // Not sure if this is cached after a split so we could have problems<a name="line.84"></a> |
| <span class="sourceLineNo">085</span> // here if a region splits while mapping<a name="line.85"></a> |
| <span class="sourceLineNo">086</span> region = this.locator.getRegionLocation(key.get()).getRegion().getStartKey();<a name="line.86"></a> |
| <span class="sourceLineNo">087</span> } catch (IOException e) {<a name="line.87"></a> |
| <span class="sourceLineNo">088</span> LOG.error(e.toString(), e);<a name="line.88"></a> |
| <span class="sourceLineNo">089</span> }<a name="line.89"></a> |
| <span class="sourceLineNo">090</span> for (int i = 0; i < this.startKeys.length; i++){<a name="line.90"></a> |
| <span class="sourceLineNo">091</span> if (Bytes.compareTo(region, this.startKeys[i]) == 0 ){<a name="line.91"></a> |
| <span class="sourceLineNo">092</span> if (i >= numPartitions-1){<a name="line.92"></a> |
| <span class="sourceLineNo">093</span> // cover if we have less reduces then regions.<a name="line.93"></a> |
| <span class="sourceLineNo">094</span> return (Integer.toString(i).hashCode()<a name="line.94"></a> |
| <span class="sourceLineNo">095</span> & Integer.MAX_VALUE) % numPartitions;<a name="line.95"></a> |
| <span class="sourceLineNo">096</span> }<a name="line.96"></a> |
| <span class="sourceLineNo">097</span> return i;<a name="line.97"></a> |
| <span class="sourceLineNo">098</span> }<a name="line.98"></a> |
| <span class="sourceLineNo">099</span> }<a name="line.99"></a> |
| <span class="sourceLineNo">100</span> // if above fails to find start key that match we need to return something<a name="line.100"></a> |
| <span class="sourceLineNo">101</span> return 0;<a name="line.101"></a> |
| <span class="sourceLineNo">102</span> }<a name="line.102"></a> |
| <span class="sourceLineNo">103</span><a name="line.103"></a> |
| <span class="sourceLineNo">104</span> /**<a name="line.104"></a> |
| <span class="sourceLineNo">105</span> * Returns the current configuration.<a name="line.105"></a> |
| <span class="sourceLineNo">106</span> *<a name="line.106"></a> |
| <span class="sourceLineNo">107</span> * @return The current configuration.<a name="line.107"></a> |
| <span class="sourceLineNo">108</span> * @see org.apache.hadoop.conf.Configurable#getConf()<a name="line.108"></a> |
| <span class="sourceLineNo">109</span> */<a name="line.109"></a> |
| <span class="sourceLineNo">110</span> @Override<a name="line.110"></a> |
| <span class="sourceLineNo">111</span> public Configuration getConf() {<a name="line.111"></a> |
| <span class="sourceLineNo">112</span> return conf;<a name="line.112"></a> |
| <span class="sourceLineNo">113</span> }<a name="line.113"></a> |
| <span class="sourceLineNo">114</span><a name="line.114"></a> |
| <span class="sourceLineNo">115</span> /**<a name="line.115"></a> |
| <span class="sourceLineNo">116</span> * Sets the configuration. This is used to determine the start keys for the<a name="line.116"></a> |
| <span class="sourceLineNo">117</span> * given table.<a name="line.117"></a> |
| <span class="sourceLineNo">118</span> *<a name="line.118"></a> |
| <span class="sourceLineNo">119</span> * @param configuration The configuration to set.<a name="line.119"></a> |
| <span class="sourceLineNo">120</span> * @see org.apache.hadoop.conf.Configurable#setConf(<a name="line.120"></a> |
| <span class="sourceLineNo">121</span> * org.apache.hadoop.conf.Configuration)<a name="line.121"></a> |
| <span class="sourceLineNo">122</span> */<a name="line.122"></a> |
| <span class="sourceLineNo">123</span> @Override<a name="line.123"></a> |
| <span class="sourceLineNo">124</span> public void setConf(Configuration configuration) {<a name="line.124"></a> |
| <span class="sourceLineNo">125</span> this.conf = HBaseConfiguration.create(configuration);<a name="line.125"></a> |
| <span class="sourceLineNo">126</span> try {<a name="line.126"></a> |
| <span class="sourceLineNo">127</span> this.connection = ConnectionFactory.createConnection(HBaseConfiguration.create(conf));<a name="line.127"></a> |
| <span class="sourceLineNo">128</span> TableName tableName = TableName.valueOf(conf.get(TableOutputFormat.OUTPUT_TABLE));<a name="line.128"></a> |
| <span class="sourceLineNo">129</span> this.locator = this.connection.getRegionLocator(tableName);<a name="line.129"></a> |
| <span class="sourceLineNo">130</span> } catch (IOException e) {<a name="line.130"></a> |
| <span class="sourceLineNo">131</span> LOG.error(e.toString(), e);<a name="line.131"></a> |
| <span class="sourceLineNo">132</span> }<a name="line.132"></a> |
| <span class="sourceLineNo">133</span> try {<a name="line.133"></a> |
| <span class="sourceLineNo">134</span> this.startKeys = this.locator.getStartKeys();<a name="line.134"></a> |
| <span class="sourceLineNo">135</span> } catch (IOException e) {<a name="line.135"></a> |
| <span class="sourceLineNo">136</span> LOG.error(e.toString(), e);<a name="line.136"></a> |
| <span class="sourceLineNo">137</span> }<a name="line.137"></a> |
| <span class="sourceLineNo">138</span> }<a name="line.138"></a> |
| <span class="sourceLineNo">139</span>}<a name="line.139"></a> |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| </pre> |
| </div> |
| </body> |
| </html> |