<HTML>
<BODY BGCOLOR="white">
<PRE>
<FONT color="green">001</FONT>    /**<a name="line.1"></a>
<FONT color="green">002</FONT>     * Licensed to the Apache Software Foundation (ASF) under one<a name="line.2"></a>
<FONT color="green">003</FONT>     * or more contributor license agreements.  See the NOTICE file<a name="line.3"></a>
<FONT color="green">004</FONT>     * distributed with this work for additional information<a name="line.4"></a>
<FONT color="green">005</FONT>     * regarding copyright ownership.  The ASF licenses this file<a name="line.5"></a>
<FONT color="green">006</FONT>     * to you under the Apache License, Version 2.0 (the<a name="line.6"></a>
<FONT color="green">007</FONT>     * "License"); you may not use this file except in compliance<a name="line.7"></a>
<FONT color="green">008</FONT>     * with the License.  You may obtain a copy of the License at<a name="line.8"></a>
<FONT color="green">009</FONT>     *<a name="line.9"></a>
<FONT color="green">010</FONT>     *     http://www.apache.org/licenses/LICENSE-2.0<a name="line.10"></a>
<FONT color="green">011</FONT>     *<a name="line.11"></a>
<FONT color="green">012</FONT>     * Unless required by applicable law or agreed to in writing, software<a name="line.12"></a>
<FONT color="green">013</FONT>     * distributed under the License is distributed on an "AS IS" BASIS,<a name="line.13"></a>
<FONT color="green">014</FONT>     * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.<a name="line.14"></a>
<FONT color="green">015</FONT>     * See the License for the specific language governing permissions and<a name="line.15"></a>
<FONT color="green">016</FONT>     * limitations under the License.<a name="line.16"></a>
<FONT color="green">017</FONT>     */<a name="line.17"></a>
<FONT color="green">018</FONT>    <a name="line.18"></a>
<FONT color="green">019</FONT>    package org.apache.hadoop.mapreduce.lib.partition;<a name="line.19"></a>
<FONT color="green">020</FONT>    <a name="line.20"></a>
<FONT color="green">021</FONT>    import java.io.IOException;<a name="line.21"></a>
<FONT color="green">022</FONT>    import java.lang.reflect.Array;<a name="line.22"></a>
<FONT color="green">023</FONT>    import java.util.ArrayList;<a name="line.23"></a>
<FONT color="green">024</FONT>    import java.util.Arrays;<a name="line.24"></a>
<FONT color="green">025</FONT>    <a name="line.25"></a>
<FONT color="green">026</FONT>    import org.apache.commons.logging.Log;<a name="line.26"></a>
<FONT color="green">027</FONT>    import org.apache.commons.logging.LogFactory;<a name="line.27"></a>
<FONT color="green">028</FONT>    import org.apache.hadoop.classification.InterfaceAudience;<a name="line.28"></a>
<FONT color="green">029</FONT>    import org.apache.hadoop.classification.InterfaceStability;<a name="line.29"></a>
<FONT color="green">030</FONT>    import org.apache.hadoop.conf.Configurable;<a name="line.30"></a>
<FONT color="green">031</FONT>    import org.apache.hadoop.conf.Configuration;<a name="line.31"></a>
<FONT color="green">032</FONT>    import org.apache.hadoop.fs.FileSystem;<a name="line.32"></a>
<FONT color="green">033</FONT>    import org.apache.hadoop.fs.Path;<a name="line.33"></a>
<FONT color="green">034</FONT>    import org.apache.hadoop.io.BinaryComparable;<a name="line.34"></a>
<FONT color="green">035</FONT>    import org.apache.hadoop.io.IOUtils;<a name="line.35"></a>
<FONT color="green">036</FONT>    import org.apache.hadoop.io.NullWritable;<a name="line.36"></a>
<FONT color="green">037</FONT>    import org.apache.hadoop.io.SequenceFile;<a name="line.37"></a>
<FONT color="green">038</FONT>    import org.apache.hadoop.io.RawComparator;<a name="line.38"></a>
<FONT color="green">039</FONT>    import org.apache.hadoop.io.WritableComparable;<a name="line.39"></a>
<FONT color="green">040</FONT>    import org.apache.hadoop.mapreduce.Job;<a name="line.40"></a>
<FONT color="green">041</FONT>    import org.apache.hadoop.mapreduce.Partitioner;<a name="line.41"></a>
<FONT color="green">042</FONT>    import org.apache.hadoop.util.ReflectionUtils;<a name="line.42"></a>
<FONT color="green">043</FONT>    <a name="line.43"></a>
<FONT color="green">044</FONT>    /**<a name="line.44"></a>
<FONT color="green">045</FONT>     * Partitioner effecting a total order by reading split points from<a name="line.45"></a>
<FONT color="green">046</FONT>     * an externally generated source.<a name="line.46"></a>
<FONT color="green">047</FONT>     */<a name="line.47"></a>
<FONT color="green">048</FONT>    @InterfaceAudience.Public<a name="line.48"></a>
<FONT color="green">049</FONT>    @InterfaceStability.Stable<a name="line.49"></a>
<FONT color="green">050</FONT>    public class TotalOrderPartitioner&lt;K extends WritableComparable&lt;?&gt;,V&gt;<a name="line.50"></a>
<FONT color="green">051</FONT>        extends Partitioner&lt;K,V&gt; implements Configurable {<a name="line.51"></a>
<FONT color="green">052</FONT>    <a name="line.52"></a>
<FONT color="green">053</FONT>      private Node partitions;<a name="line.53"></a>
<FONT color="green">054</FONT>      public static final String DEFAULT_PATH = "_partition.lst";<a name="line.54"></a>
<FONT color="green">055</FONT>      public static final String PARTITIONER_PATH = <a name="line.55"></a>
<FONT color="green">056</FONT>        "mapreduce.totalorderpartitioner.path";<a name="line.56"></a>
<FONT color="green">057</FONT>      public static final String MAX_TRIE_DEPTH = <a name="line.57"></a>
<FONT color="green">058</FONT>        "mapreduce.totalorderpartitioner.trie.maxdepth"; <a name="line.58"></a>
<FONT color="green">059</FONT>      public static final String NATURAL_ORDER = <a name="line.59"></a>
<FONT color="green">060</FONT>        "mapreduce.totalorderpartitioner.naturalorder";<a name="line.60"></a>
<FONT color="green">061</FONT>      Configuration conf;<a name="line.61"></a>
<FONT color="green">062</FONT>      private static final Log LOG = LogFactory.getLog(TotalOrderPartitioner.class);<a name="line.62"></a>
<FONT color="green">063</FONT>    <a name="line.63"></a>
<FONT color="green">064</FONT>      public TotalOrderPartitioner() { }<a name="line.64"></a>
<FONT color="green">065</FONT>    <a name="line.65"></a>
<FONT color="green">066</FONT>      /**<a name="line.66"></a>
<FONT color="green">067</FONT>       * Read in the partition file and build indexing data structures.<a name="line.67"></a>
<FONT color="green">068</FONT>       * If the keytype is {@link org.apache.hadoop.io.BinaryComparable} and<a name="line.68"></a>
<FONT color="green">069</FONT>       * &lt;tt&gt;total.order.partitioner.natural.order&lt;/tt&gt; is not false, a trie<a name="line.69"></a>
<FONT color="green">070</FONT>       * of the first &lt;tt&gt;total.order.partitioner.max.trie.depth&lt;/tt&gt;(2) + 1 bytes<a name="line.70"></a>
<FONT color="green">071</FONT>       * will be built. Otherwise, keys will be located using a binary search of<a name="line.71"></a>
<FONT color="green">072</FONT>       * the partition keyset using the {@link org.apache.hadoop.io.RawComparator}<a name="line.72"></a>
<FONT color="green">073</FONT>       * defined for this job. The input file must be sorted with the same<a name="line.73"></a>
<FONT color="green">074</FONT>       * comparator and contain {@link Job#getNumReduceTasks()} - 1 keys.<a name="line.74"></a>
<FONT color="green">075</FONT>       */<a name="line.75"></a>
<FONT color="green">076</FONT>      @SuppressWarnings("unchecked") // keytype from conf not static<a name="line.76"></a>
<FONT color="green">077</FONT>      public void setConf(Configuration conf) {<a name="line.77"></a>
<FONT color="green">078</FONT>        try {<a name="line.78"></a>
<FONT color="green">079</FONT>          this.conf = conf;<a name="line.79"></a>
<FONT color="green">080</FONT>          String parts = getPartitionFile(conf);<a name="line.80"></a>
<FONT color="green">081</FONT>          final Path partFile = new Path(parts);<a name="line.81"></a>
<FONT color="green">082</FONT>          final FileSystem fs = (DEFAULT_PATH.equals(parts))<a name="line.82"></a>
<FONT color="green">083</FONT>            ? FileSystem.getLocal(conf)     // assume in DistributedCache<a name="line.83"></a>
<FONT color="green">084</FONT>            : partFile.getFileSystem(conf);<a name="line.84"></a>
<FONT color="green">085</FONT>    <a name="line.85"></a>
<FONT color="green">086</FONT>          Job job = new Job(conf);<a name="line.86"></a>
<FONT color="green">087</FONT>          Class&lt;K&gt; keyClass = (Class&lt;K&gt;)job.getMapOutputKeyClass();<a name="line.87"></a>
<FONT color="green">088</FONT>          K[] splitPoints = readPartitions(fs, partFile, keyClass, conf);<a name="line.88"></a>
<FONT color="green">089</FONT>          if (splitPoints.length != job.getNumReduceTasks() - 1) {<a name="line.89"></a>
<FONT color="green">090</FONT>            throw new IOException("Wrong number of partitions in keyset");<a name="line.90"></a>
<FONT color="green">091</FONT>          }<a name="line.91"></a>
<FONT color="green">092</FONT>          RawComparator&lt;K&gt; comparator =<a name="line.92"></a>
<FONT color="green">093</FONT>            (RawComparator&lt;K&gt;) job.getSortComparator();<a name="line.93"></a>
<FONT color="green">094</FONT>          for (int i = 0; i &lt; splitPoints.length - 1; ++i) {<a name="line.94"></a>
<FONT color="green">095</FONT>            if (comparator.compare(splitPoints[i], splitPoints[i+1]) &gt;= 0) {<a name="line.95"></a>
<FONT color="green">096</FONT>              throw new IOException("Split points are out of order");<a name="line.96"></a>
<FONT color="green">097</FONT>            }<a name="line.97"></a>
<FONT color="green">098</FONT>          }<a name="line.98"></a>
<FONT color="green">099</FONT>          boolean natOrder =<a name="line.99"></a>
<FONT color="green">100</FONT>            conf.getBoolean(NATURAL_ORDER, true);<a name="line.100"></a>
<FONT color="green">101</FONT>          if (natOrder &amp;&amp; BinaryComparable.class.isAssignableFrom(keyClass)) {<a name="line.101"></a>
<FONT color="green">102</FONT>            partitions = buildTrie((BinaryComparable[])splitPoints, 0,<a name="line.102"></a>
<FONT color="green">103</FONT>                splitPoints.length, new byte[0],<a name="line.103"></a>
<FONT color="green">104</FONT>                // Now that blocks of identical splitless trie nodes are <a name="line.104"></a>
<FONT color="green">105</FONT>                // represented reentrantly, and we develop a leaf for any trie<a name="line.105"></a>
<FONT color="green">106</FONT>                // node with only one split point, the only reason for a depth<a name="line.106"></a>
<FONT color="green">107</FONT>                // limit is to refute stack overflow or bloat in the pathological<a name="line.107"></a>
<FONT color="green">108</FONT>                // case where the split points are long and mostly look like bytes <a name="line.108"></a>
<FONT color="green">109</FONT>                // iii...iixii...iii   .  Therefore, we make the default depth<a name="line.109"></a>
<FONT color="green">110</FONT>                // limit large but not huge.<a name="line.110"></a>
<FONT color="green">111</FONT>                conf.getInt(MAX_TRIE_DEPTH, 200));<a name="line.111"></a>
<FONT color="green">112</FONT>          } else {<a name="line.112"></a>
<FONT color="green">113</FONT>            partitions = new BinarySearchNode(splitPoints, comparator);<a name="line.113"></a>
<FONT color="green">114</FONT>          }<a name="line.114"></a>
<FONT color="green">115</FONT>        } catch (IOException e) {<a name="line.115"></a>
<FONT color="green">116</FONT>          throw new IllegalArgumentException("Can't read partitions file", e);<a name="line.116"></a>
<FONT color="green">117</FONT>        }<a name="line.117"></a>
<FONT color="green">118</FONT>      }<a name="line.118"></a>
<FONT color="green">119</FONT>    <a name="line.119"></a>
<FONT color="green">120</FONT>      public Configuration getConf() {<a name="line.120"></a>
<FONT color="green">121</FONT>        return conf;<a name="line.121"></a>
<FONT color="green">122</FONT>      }<a name="line.122"></a>
<FONT color="green">123</FONT>      <a name="line.123"></a>
<FONT color="green">124</FONT>      // by construction, we know if our keytype<a name="line.124"></a>
<FONT color="green">125</FONT>      @SuppressWarnings("unchecked") // is memcmp-able and uses the trie<a name="line.125"></a>
<FONT color="green">126</FONT>      public int getPartition(K key, V value, int numPartitions) {<a name="line.126"></a>
<FONT color="green">127</FONT>        return partitions.findPartition(key);<a name="line.127"></a>
<FONT color="green">128</FONT>      }<a name="line.128"></a>
<FONT color="green">129</FONT>    <a name="line.129"></a>
<FONT color="green">130</FONT>      /**<a name="line.130"></a>
<FONT color="green">131</FONT>       * Set the path to the SequenceFile storing the sorted partition keyset.<a name="line.131"></a>
<FONT color="green">132</FONT>       * It must be the case that for &lt;tt&gt;R&lt;/tt&gt; reduces, there are &lt;tt&gt;R-1&lt;/tt&gt;<a name="line.132"></a>
<FONT color="green">133</FONT>       * keys in the SequenceFile.<a name="line.133"></a>
<FONT color="green">134</FONT>       */<a name="line.134"></a>
<FONT color="green">135</FONT>      public static void setPartitionFile(Configuration conf, Path p) {<a name="line.135"></a>
<FONT color="green">136</FONT>        conf.set(PARTITIONER_PATH, p.toString());<a name="line.136"></a>
<FONT color="green">137</FONT>      }<a name="line.137"></a>
<FONT color="green">138</FONT>    <a name="line.138"></a>
<FONT color="green">139</FONT>      /**<a name="line.139"></a>
<FONT color="green">140</FONT>       * Get the path to the SequenceFile storing the sorted partition keyset.<a name="line.140"></a>
<FONT color="green">141</FONT>       * @see #setPartitionFile(Configuration, Path)<a name="line.141"></a>
<FONT color="green">142</FONT>       */<a name="line.142"></a>
<FONT color="green">143</FONT>      public static String getPartitionFile(Configuration conf) {<a name="line.143"></a>
<FONT color="green">144</FONT>        return conf.get(PARTITIONER_PATH, DEFAULT_PATH);<a name="line.144"></a>
<FONT color="green">145</FONT>      }<a name="line.145"></a>
<FONT color="green">146</FONT>    <a name="line.146"></a>
<FONT color="green">147</FONT>      /**<a name="line.147"></a>
<FONT color="green">148</FONT>       * Interface to the partitioner to locate a key in the partition keyset.<a name="line.148"></a>
<FONT color="green">149</FONT>       */<a name="line.149"></a>
<FONT color="green">150</FONT>      interface Node&lt;T&gt; {<a name="line.150"></a>
<FONT color="green">151</FONT>        /**<a name="line.151"></a>
<FONT color="green">152</FONT>         * Locate partition in keyset K, st [Ki..Ki+1) defines a partition,<a name="line.152"></a>
<FONT color="green">153</FONT>         * with implicit K0 = -inf, Kn = +inf, and |K| = #partitions - 1.<a name="line.153"></a>
<FONT color="green">154</FONT>         */<a name="line.154"></a>
<FONT color="green">155</FONT>        int findPartition(T key);<a name="line.155"></a>
<FONT color="green">156</FONT>      }<a name="line.156"></a>
<FONT color="green">157</FONT>    <a name="line.157"></a>
<FONT color="green">158</FONT>      /**<a name="line.158"></a>
<FONT color="green">159</FONT>       * Base class for trie nodes. If the keytype is memcomp-able, this builds<a name="line.159"></a>
<FONT color="green">160</FONT>       * tries of the first &lt;tt&gt;total.order.partitioner.max.trie.depth&lt;/tt&gt;<a name="line.160"></a>
<FONT color="green">161</FONT>       * bytes.<a name="line.161"></a>
<FONT color="green">162</FONT>       */<a name="line.162"></a>
<FONT color="green">163</FONT>      static abstract class TrieNode implements Node&lt;BinaryComparable&gt; {<a name="line.163"></a>
<FONT color="green">164</FONT>        private final int level;<a name="line.164"></a>
<FONT color="green">165</FONT>        TrieNode(int level) {<a name="line.165"></a>
<FONT color="green">166</FONT>          this.level = level;<a name="line.166"></a>
<FONT color="green">167</FONT>        }<a name="line.167"></a>
<FONT color="green">168</FONT>        int getLevel() {<a name="line.168"></a>
<FONT color="green">169</FONT>          return level;<a name="line.169"></a>
<FONT color="green">170</FONT>        }<a name="line.170"></a>
<FONT color="green">171</FONT>      }<a name="line.171"></a>
<FONT color="green">172</FONT>    <a name="line.172"></a>
<FONT color="green">173</FONT>      /**<a name="line.173"></a>
<FONT color="green">174</FONT>       * For types that are not {@link org.apache.hadoop.io.BinaryComparable} or<a name="line.174"></a>
<FONT color="green">175</FONT>       * where disabled by &lt;tt&gt;total.order.partitioner.natural.order&lt;/tt&gt;,<a name="line.175"></a>
<FONT color="green">176</FONT>       * search the partition keyset with a binary search.<a name="line.176"></a>
<FONT color="green">177</FONT>       */<a name="line.177"></a>
<FONT color="green">178</FONT>      class BinarySearchNode implements Node&lt;K&gt; {<a name="line.178"></a>
<FONT color="green">179</FONT>        private final K[] splitPoints;<a name="line.179"></a>
<FONT color="green">180</FONT>        private final RawComparator&lt;K&gt; comparator;<a name="line.180"></a>
<FONT color="green">181</FONT>        BinarySearchNode(K[] splitPoints, RawComparator&lt;K&gt; comparator) {<a name="line.181"></a>
<FONT color="green">182</FONT>          this.splitPoints = splitPoints;<a name="line.182"></a>
<FONT color="green">183</FONT>          this.comparator = comparator;<a name="line.183"></a>
<FONT color="green">184</FONT>        }<a name="line.184"></a>
<FONT color="green">185</FONT>        public int findPartition(K key) {<a name="line.185"></a>
<FONT color="green">186</FONT>          final int pos = Arrays.binarySearch(splitPoints, key, comparator) + 1;<a name="line.186"></a>
<FONT color="green">187</FONT>          return (pos &lt; 0) ? -pos : pos;<a name="line.187"></a>
<FONT color="green">188</FONT>        }<a name="line.188"></a>
<FONT color="green">189</FONT>      }<a name="line.189"></a>
<FONT color="green">190</FONT>    <a name="line.190"></a>
<FONT color="green">191</FONT>      /**<a name="line.191"></a>
<FONT color="green">192</FONT>       * An inner trie node that contains 256 children based on the next<a name="line.192"></a>
<FONT color="green">193</FONT>       * character.<a name="line.193"></a>
<FONT color="green">194</FONT>       */<a name="line.194"></a>
<FONT color="green">195</FONT>      class InnerTrieNode extends TrieNode {<a name="line.195"></a>
<FONT color="green">196</FONT>        private TrieNode[] child = new TrieNode[256];<a name="line.196"></a>
<FONT color="green">197</FONT>    <a name="line.197"></a>
<FONT color="green">198</FONT>        InnerTrieNode(int level) {<a name="line.198"></a>
<FONT color="green">199</FONT>          super(level);<a name="line.199"></a>
<FONT color="green">200</FONT>        }<a name="line.200"></a>
<FONT color="green">201</FONT>        public int findPartition(BinaryComparable key) {<a name="line.201"></a>
<FONT color="green">202</FONT>          int level = getLevel();<a name="line.202"></a>
<FONT color="green">203</FONT>          if (key.getLength() &lt;= level) {<a name="line.203"></a>
<FONT color="green">204</FONT>            return child[0].findPartition(key);<a name="line.204"></a>
<FONT color="green">205</FONT>          }<a name="line.205"></a>
<FONT color="green">206</FONT>          return child[0xFF &amp; key.getBytes()[level]].findPartition(key);<a name="line.206"></a>
<FONT color="green">207</FONT>        }<a name="line.207"></a>
<FONT color="green">208</FONT>      }<a name="line.208"></a>
<FONT color="green">209</FONT>      <a name="line.209"></a>
<FONT color="green">210</FONT>      /**<a name="line.210"></a>
<FONT color="green">211</FONT>       * @param level        the tree depth at this node<a name="line.211"></a>
<FONT color="green">212</FONT>       * @param splitPoints  the full split point vector, which holds<a name="line.212"></a>
<FONT color="green">213</FONT>       *                     the split point or points this leaf node<a name="line.213"></a>
<FONT color="green">214</FONT>       *                     should contain<a name="line.214"></a>
<FONT color="green">215</FONT>       * @param lower        first INcluded element of splitPoints<a name="line.215"></a>
<FONT color="green">216</FONT>       * @param upper        first EXcluded element of splitPoints<a name="line.216"></a>
<FONT color="green">217</FONT>       * @return  a leaf node.  They come in three kinds: no split points <a name="line.217"></a>
<FONT color="green">218</FONT>       *          [and the findParttion returns a canned index], one split<a name="line.218"></a>
<FONT color="green">219</FONT>       *          point [and we compare with a single comparand], or more<a name="line.219"></a>
<FONT color="green">220</FONT>       *          than one [and we do a binary search].  The last case is<a name="line.220"></a>
<FONT color="green">221</FONT>       *          rare.<a name="line.221"></a>
<FONT color="green">222</FONT>       */<a name="line.222"></a>
<FONT color="green">223</FONT>      private TrieNode LeafTrieNodeFactory<a name="line.223"></a>
<FONT color="green">224</FONT>                 (int level, BinaryComparable[] splitPoints, int lower, int upper) {<a name="line.224"></a>
<FONT color="green">225</FONT>          switch (upper - lower) {<a name="line.225"></a>
<FONT color="green">226</FONT>          case 0:<a name="line.226"></a>
<FONT color="green">227</FONT>              return new UnsplitTrieNode(level, lower);<a name="line.227"></a>
<FONT color="green">228</FONT>              <a name="line.228"></a>
<FONT color="green">229</FONT>          case 1:<a name="line.229"></a>
<FONT color="green">230</FONT>              return new SinglySplitTrieNode(level, splitPoints, lower);<a name="line.230"></a>
<FONT color="green">231</FONT>              <a name="line.231"></a>
<FONT color="green">232</FONT>          default:<a name="line.232"></a>
<FONT color="green">233</FONT>              return new LeafTrieNode(level, splitPoints, lower, upper);<a name="line.233"></a>
<FONT color="green">234</FONT>          }<a name="line.234"></a>
<FONT color="green">235</FONT>      }<a name="line.235"></a>
<FONT color="green">236</FONT>    <a name="line.236"></a>
<FONT color="green">237</FONT>      /**<a name="line.237"></a>
<FONT color="green">238</FONT>       * A leaf trie node that scans for the key between lower..upper.<a name="line.238"></a>
<FONT color="green">239</FONT>       * <a name="line.239"></a>
<FONT color="green">240</FONT>       * We don't generate many of these now, since we usually continue trie-ing <a name="line.240"></a>
<FONT color="green">241</FONT>       * when more than one split point remains at this level. and we make different<a name="line.241"></a>
<FONT color="green">242</FONT>       * objects for nodes with 0 or 1 split point.<a name="line.242"></a>
<FONT color="green">243</FONT>       */<a name="line.243"></a>
<FONT color="green">244</FONT>      private class LeafTrieNode extends TrieNode {<a name="line.244"></a>
<FONT color="green">245</FONT>        final int lower;<a name="line.245"></a>
<FONT color="green">246</FONT>        final int upper;<a name="line.246"></a>
<FONT color="green">247</FONT>        final BinaryComparable[] splitPoints;<a name="line.247"></a>
<FONT color="green">248</FONT>        LeafTrieNode(int level, BinaryComparable[] splitPoints, int lower, int upper) {<a name="line.248"></a>
<FONT color="green">249</FONT>          super(level);<a name="line.249"></a>
<FONT color="green">250</FONT>          this.lower = lower;<a name="line.250"></a>
<FONT color="green">251</FONT>          this.upper = upper;<a name="line.251"></a>
<FONT color="green">252</FONT>          this.splitPoints = splitPoints;<a name="line.252"></a>
<FONT color="green">253</FONT>        }<a name="line.253"></a>
<FONT color="green">254</FONT>        public int findPartition(BinaryComparable key) {<a name="line.254"></a>
<FONT color="green">255</FONT>          final int pos = Arrays.binarySearch(splitPoints, lower, upper, key) + 1;<a name="line.255"></a>
<FONT color="green">256</FONT>          return (pos &lt; 0) ? -pos : pos;<a name="line.256"></a>
<FONT color="green">257</FONT>        }<a name="line.257"></a>
<FONT color="green">258</FONT>      }<a name="line.258"></a>
<FONT color="green">259</FONT>      <a name="line.259"></a>
<FONT color="green">260</FONT>      private class UnsplitTrieNode extends TrieNode {<a name="line.260"></a>
<FONT color="green">261</FONT>          final int result;<a name="line.261"></a>
<FONT color="green">262</FONT>          <a name="line.262"></a>
<FONT color="green">263</FONT>          UnsplitTrieNode(int level, int value) {<a name="line.263"></a>
<FONT color="green">264</FONT>              super(level);<a name="line.264"></a>
<FONT color="green">265</FONT>              this.result = value;<a name="line.265"></a>
<FONT color="green">266</FONT>          }<a name="line.266"></a>
<FONT color="green">267</FONT>          <a name="line.267"></a>
<FONT color="green">268</FONT>          public int findPartition(BinaryComparable key) {<a name="line.268"></a>
<FONT color="green">269</FONT>              return result;<a name="line.269"></a>
<FONT color="green">270</FONT>          }<a name="line.270"></a>
<FONT color="green">271</FONT>      }<a name="line.271"></a>
<FONT color="green">272</FONT>      <a name="line.272"></a>
<FONT color="green">273</FONT>      private class SinglySplitTrieNode extends TrieNode {<a name="line.273"></a>
<FONT color="green">274</FONT>          final int               lower;<a name="line.274"></a>
<FONT color="green">275</FONT>          final BinaryComparable  mySplitPoint;<a name="line.275"></a>
<FONT color="green">276</FONT>          <a name="line.276"></a>
<FONT color="green">277</FONT>          SinglySplitTrieNode(int level, BinaryComparable[] splitPoints, int lower) {<a name="line.277"></a>
<FONT color="green">278</FONT>              super(level);<a name="line.278"></a>
<FONT color="green">279</FONT>              this.lower = lower;<a name="line.279"></a>
<FONT color="green">280</FONT>              this.mySplitPoint = splitPoints[lower];<a name="line.280"></a>
<FONT color="green">281</FONT>          }<a name="line.281"></a>
<FONT color="green">282</FONT>          <a name="line.282"></a>
<FONT color="green">283</FONT>          public int findPartition(BinaryComparable key) {<a name="line.283"></a>
<FONT color="green">284</FONT>              return lower + (key.compareTo(mySplitPoint) &lt; 0 ? 0 : 1);<a name="line.284"></a>
<FONT color="green">285</FONT>          }<a name="line.285"></a>
<FONT color="green">286</FONT>      }<a name="line.286"></a>
<FONT color="green">287</FONT>    <a name="line.287"></a>
<FONT color="green">288</FONT>    <a name="line.288"></a>
<FONT color="green">289</FONT>      /**<a name="line.289"></a>
<FONT color="green">290</FONT>       * Read the cut points from the given IFile.<a name="line.290"></a>
<FONT color="green">291</FONT>       * @param fs The file system<a name="line.291"></a>
<FONT color="green">292</FONT>       * @param p The path to read<a name="line.292"></a>
<FONT color="green">293</FONT>       * @param keyClass The map output key class<a name="line.293"></a>
<FONT color="green">294</FONT>       * @param job The job config<a name="line.294"></a>
<FONT color="green">295</FONT>       * @throws IOException<a name="line.295"></a>
<FONT color="green">296</FONT>       */<a name="line.296"></a>
<FONT color="green">297</FONT>                                     // matching key types enforced by passing in<a name="line.297"></a>
<FONT color="green">298</FONT>      @SuppressWarnings("unchecked") // map output key class<a name="line.298"></a>
<FONT color="green">299</FONT>      private K[] readPartitions(FileSystem fs, Path p, Class&lt;K&gt; keyClass,<a name="line.299"></a>
<FONT color="green">300</FONT>          Configuration conf) throws IOException {<a name="line.300"></a>
<FONT color="green">301</FONT>        SequenceFile.Reader reader = new SequenceFile.Reader(fs, p, conf);<a name="line.301"></a>
<FONT color="green">302</FONT>        ArrayList&lt;K&gt; parts = new ArrayList&lt;K&gt;();<a name="line.302"></a>
<FONT color="green">303</FONT>        K key = ReflectionUtils.newInstance(keyClass, conf);<a name="line.303"></a>
<FONT color="green">304</FONT>        NullWritable value = NullWritable.get();<a name="line.304"></a>
<FONT color="green">305</FONT>        try {<a name="line.305"></a>
<FONT color="green">306</FONT>          while (reader.next(key, value)) {<a name="line.306"></a>
<FONT color="green">307</FONT>            parts.add(key);<a name="line.307"></a>
<FONT color="green">308</FONT>            key = ReflectionUtils.newInstance(keyClass, conf);<a name="line.308"></a>
<FONT color="green">309</FONT>          }<a name="line.309"></a>
<FONT color="green">310</FONT>          reader.close();<a name="line.310"></a>
<FONT color="green">311</FONT>          reader = null;<a name="line.311"></a>
<FONT color="green">312</FONT>        } finally {<a name="line.312"></a>
<FONT color="green">313</FONT>          IOUtils.cleanup(LOG, reader);<a name="line.313"></a>
<FONT color="green">314</FONT>        }<a name="line.314"></a>
<FONT color="green">315</FONT>        return parts.toArray((K[])Array.newInstance(keyClass, parts.size()));<a name="line.315"></a>
<FONT color="green">316</FONT>      }<a name="line.316"></a>
<FONT color="green">317</FONT>      <a name="line.317"></a>
<FONT color="green">318</FONT>      /**<a name="line.318"></a>
<FONT color="green">319</FONT>       * <a name="line.319"></a>
<FONT color="green">320</FONT>       * This object contains a TrieNodeRef if there is such a thing that<a name="line.320"></a>
<FONT color="green">321</FONT>       * can be repeated.  Two adjacent trie node slots that contain no <a name="line.321"></a>
<FONT color="green">322</FONT>       * split points can be filled with the same trie node, even if they<a name="line.322"></a>
<FONT color="green">323</FONT>       * are not on the same level.  See buildTreeRec, below.<a name="line.323"></a>
<FONT color="green">324</FONT>       *<a name="line.324"></a>
<FONT color="green">325</FONT>       */  <a name="line.325"></a>
<FONT color="green">326</FONT>      private class CarriedTrieNodeRef<a name="line.326"></a>
<FONT color="green">327</FONT>      {<a name="line.327"></a>
<FONT color="green">328</FONT>          TrieNode   content;<a name="line.328"></a>
<FONT color="green">329</FONT>          <a name="line.329"></a>
<FONT color="green">330</FONT>          CarriedTrieNodeRef() {<a name="line.330"></a>
<FONT color="green">331</FONT>              content = null;<a name="line.331"></a>
<FONT color="green">332</FONT>          }<a name="line.332"></a>
<FONT color="green">333</FONT>      }<a name="line.333"></a>
<FONT color="green">334</FONT>    <a name="line.334"></a>
<FONT color="green">335</FONT>      <a name="line.335"></a>
<FONT color="green">336</FONT>      /**<a name="line.336"></a>
<FONT color="green">337</FONT>       * Given a sorted set of cut points, build a trie that will find the correct<a name="line.337"></a>
<FONT color="green">338</FONT>       * partition quickly.<a name="line.338"></a>
<FONT color="green">339</FONT>       * @param splits the list of cut points<a name="line.339"></a>
<FONT color="green">340</FONT>       * @param lower the lower bound of partitions 0..numPartitions-1<a name="line.340"></a>
<FONT color="green">341</FONT>       * @param upper the upper bound of partitions 0..numPartitions-1<a name="line.341"></a>
<FONT color="green">342</FONT>       * @param prefix the prefix that we have already checked against<a name="line.342"></a>
<FONT color="green">343</FONT>       * @param maxDepth the maximum depth we will build a trie for<a name="line.343"></a>
<FONT color="green">344</FONT>       * @return the trie node that will divide the splits correctly<a name="line.344"></a>
<FONT color="green">345</FONT>       */<a name="line.345"></a>
<FONT color="green">346</FONT>      private TrieNode buildTrie(BinaryComparable[] splits, int lower,<a name="line.346"></a>
<FONT color="green">347</FONT>              int upper, byte[] prefix, int maxDepth) {<a name="line.347"></a>
<FONT color="green">348</FONT>          return buildTrieRec<a name="line.348"></a>
<FONT color="green">349</FONT>                   (splits, lower, upper, prefix, maxDepth, new CarriedTrieNodeRef());<a name="line.349"></a>
<FONT color="green">350</FONT>      }<a name="line.350"></a>
<FONT color="green">351</FONT>      <a name="line.351"></a>
<FONT color="green">352</FONT>      /**<a name="line.352"></a>
<FONT color="green">353</FONT>       * This is the core of buildTrie.  The interface, and stub, above, just adds<a name="line.353"></a>
<FONT color="green">354</FONT>       * an empty CarriedTrieNodeRef.  <a name="line.354"></a>
<FONT color="green">355</FONT>       * <a name="line.355"></a>
<FONT color="green">356</FONT>       * We build trie nodes in depth first order, which is also in key space<a name="line.356"></a>
<FONT color="green">357</FONT>       * order.  Every leaf node is referenced as a slot in a parent internal<a name="line.357"></a>
<FONT color="green">358</FONT>       * node.  If two adjacent slots [in the DFO] hold leaf nodes that have<a name="line.358"></a>
<FONT color="green">359</FONT>       * no split point, then they are not separated by a split point either, <a name="line.359"></a>
<FONT color="green">360</FONT>       * because there's no place in key space for that split point to exist.<a name="line.360"></a>
<FONT color="green">361</FONT>       * <a name="line.361"></a>
<FONT color="green">362</FONT>       * When that happens, the leaf nodes would be semantically identical, and<a name="line.362"></a>
<FONT color="green">363</FONT>       * we reuse the object.  A single CarriedTrieNodeRef "ref" lives for the <a name="line.363"></a>
<FONT color="green">364</FONT>       * duration of the tree-walk.  ref carries a potentially reusable, unsplit<a name="line.364"></a>
<FONT color="green">365</FONT>       * leaf node for such reuse until a leaf node with a split arises, which <a name="line.365"></a>
<FONT color="green">366</FONT>       * breaks the chain until we need to make a new unsplit leaf node.<a name="line.366"></a>
<FONT color="green">367</FONT>       * <a name="line.367"></a>
<FONT color="green">368</FONT>       * Note that this use of CarriedTrieNodeRef means that for internal nodes, <a name="line.368"></a>
<FONT color="green">369</FONT>       * for internal nodes if this code is modified in any way we still need <a name="line.369"></a>
<FONT color="green">370</FONT>       * to make or fill in the subnodes in key space order.<a name="line.370"></a>
<FONT color="green">371</FONT>       */<a name="line.371"></a>
<FONT color="green">372</FONT>      private TrieNode buildTrieRec(BinaryComparable[] splits, int lower,<a name="line.372"></a>
<FONT color="green">373</FONT>          int upper, byte[] prefix, int maxDepth, CarriedTrieNodeRef ref) {<a name="line.373"></a>
<FONT color="green">374</FONT>        final int depth = prefix.length;<a name="line.374"></a>
<FONT color="green">375</FONT>        // We generate leaves for a single split point as well as for <a name="line.375"></a>
<FONT color="green">376</FONT>        // no split points.<a name="line.376"></a>
<FONT color="green">377</FONT>        if (depth &gt;= maxDepth || lower &gt;= upper - 1) {<a name="line.377"></a>
<FONT color="green">378</FONT>            // If we have two consecutive requests for an unsplit trie node, we<a name="line.378"></a>
<FONT color="green">379</FONT>            // can deliver the same one the second time.<a name="line.379"></a>
<FONT color="green">380</FONT>            if (lower == upper &amp;&amp; ref.content != null) {<a name="line.380"></a>
<FONT color="green">381</FONT>                return ref.content;<a name="line.381"></a>
<FONT color="green">382</FONT>            }<a name="line.382"></a>
<FONT color="green">383</FONT>            TrieNode  result = LeafTrieNodeFactory(depth, splits, lower, upper);<a name="line.383"></a>
<FONT color="green">384</FONT>            ref.content = lower == upper ? result : null;<a name="line.384"></a>
<FONT color="green">385</FONT>            return result;<a name="line.385"></a>
<FONT color="green">386</FONT>        }<a name="line.386"></a>
<FONT color="green">387</FONT>        InnerTrieNode result = new InnerTrieNode(depth);<a name="line.387"></a>
<FONT color="green">388</FONT>        byte[] trial = Arrays.copyOf(prefix, prefix.length + 1);<a name="line.388"></a>
<FONT color="green">389</FONT>        // append an extra byte on to the prefix<a name="line.389"></a>
<FONT color="green">390</FONT>        int         currentBound = lower;<a name="line.390"></a>
<FONT color="green">391</FONT>        for(int ch = 0; ch &lt; 0xFF; ++ch) {<a name="line.391"></a>
<FONT color="green">392</FONT>          trial[depth] = (byte) (ch + 1);<a name="line.392"></a>
<FONT color="green">393</FONT>          lower = currentBound;<a name="line.393"></a>
<FONT color="green">394</FONT>          while (currentBound &lt; upper) {<a name="line.394"></a>
<FONT color="green">395</FONT>            if (splits[currentBound].compareTo(trial, 0, trial.length) &gt;= 0) {<a name="line.395"></a>
<FONT color="green">396</FONT>              break;<a name="line.396"></a>
<FONT color="green">397</FONT>            }<a name="line.397"></a>
<FONT color="green">398</FONT>            currentBound += 1;<a name="line.398"></a>
<FONT color="green">399</FONT>          }<a name="line.399"></a>
<FONT color="green">400</FONT>          trial[depth] = (byte) ch;<a name="line.400"></a>
<FONT color="green">401</FONT>          result.child[0xFF &amp; ch]<a name="line.401"></a>
<FONT color="green">402</FONT>                       = buildTrieRec(splits, lower, currentBound, trial, maxDepth, ref);<a name="line.402"></a>
<FONT color="green">403</FONT>        }<a name="line.403"></a>
<FONT color="green">404</FONT>        // pick up the rest<a name="line.404"></a>
<FONT color="green">405</FONT>        trial[depth] = (byte)0xFF;<a name="line.405"></a>
<FONT color="green">406</FONT>        result.child[0xFF] <a name="line.406"></a>
<FONT color="green">407</FONT>                     = buildTrieRec(splits, lower, currentBound, trial, maxDepth, ref);<a name="line.407"></a>
<FONT color="green">408</FONT>        <a name="line.408"></a>
<FONT color="green">409</FONT>        return result;<a name="line.409"></a>
<FONT color="green">410</FONT>      }<a name="line.410"></a>
<FONT color="green">411</FONT>    }<a name="line.411"></a>




























































</PRE>
</BODY>
</HTML>
