blob: 0a380b7fca95f380a0b1581b920a27a0bceb0e68 [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> * Licensed to the Apache Software Foundation (ASF) under one<a name="line.2"></a>
<span class="sourceLineNo">003</span> * or more contributor license agreements. See the NOTICE file<a name="line.3"></a>
<span class="sourceLineNo">004</span> * distributed with this work for additional information<a name="line.4"></a>
<span class="sourceLineNo">005</span> * regarding copyright ownership. The ASF licenses this file<a name="line.5"></a>
<span class="sourceLineNo">006</span> * to you under the Apache License, Version 2.0 (the<a name="line.6"></a>
<span class="sourceLineNo">007</span> * "License"); you may not use this file except in compliance<a name="line.7"></a>
<span class="sourceLineNo">008</span> * with the License. You may obtain a copy of the License at<a name="line.8"></a>
<span class="sourceLineNo">009</span> *<a name="line.9"></a>
<span class="sourceLineNo">010</span> * http://www.apache.org/licenses/LICENSE-2.0<a name="line.10"></a>
<span class="sourceLineNo">011</span> *<a name="line.11"></a>
<span class="sourceLineNo">012</span> * Unless required by applicable law or agreed to in writing, software<a name="line.12"></a>
<span class="sourceLineNo">013</span> * distributed under the License is distributed on an "AS IS" BASIS,<a name="line.13"></a>
<span class="sourceLineNo">014</span> * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.<a name="line.14"></a>
<span class="sourceLineNo">015</span> * See the License for the specific language governing permissions and<a name="line.15"></a>
<span class="sourceLineNo">016</span> * limitations under the License.<a name="line.16"></a>
<span class="sourceLineNo">017</span> */<a name="line.17"></a>
<span class="sourceLineNo">018</span>package org.apache.hadoop.hbase.mapred;<a name="line.18"></a>
<span class="sourceLineNo">019</span><a name="line.19"></a>
<span class="sourceLineNo">020</span>import java.io.IOException;<a name="line.20"></a>
<span class="sourceLineNo">021</span>import java.util.ArrayList;<a name="line.21"></a>
<span class="sourceLineNo">022</span>import org.apache.hadoop.hbase.Cell;<a name="line.22"></a>
<span class="sourceLineNo">023</span>import org.apache.hadoop.hbase.CellUtil;<a name="line.23"></a>
<span class="sourceLineNo">024</span>import org.apache.hadoop.hbase.client.Result;<a name="line.24"></a>
<span class="sourceLineNo">025</span>import org.apache.hadoop.hbase.io.ImmutableBytesWritable;<a name="line.25"></a>
<span class="sourceLineNo">026</span>import org.apache.hadoop.hbase.util.Bytes;<a name="line.26"></a>
<span class="sourceLineNo">027</span>import org.apache.hadoop.mapred.JobConf;<a name="line.27"></a>
<span class="sourceLineNo">028</span>import org.apache.hadoop.mapred.MapReduceBase;<a name="line.28"></a>
<span class="sourceLineNo">029</span>import org.apache.hadoop.mapred.OutputCollector;<a name="line.29"></a>
<span class="sourceLineNo">030</span>import org.apache.hadoop.mapred.Reporter;<a name="line.30"></a>
<span class="sourceLineNo">031</span>import org.apache.yetus.audience.InterfaceAudience;<a name="line.31"></a>
<span class="sourceLineNo">032</span><a name="line.32"></a>
<span class="sourceLineNo">033</span>/**<a name="line.33"></a>
<span class="sourceLineNo">034</span> * Extract grouping columns from input record<a name="line.34"></a>
<span class="sourceLineNo">035</span> */<a name="line.35"></a>
<span class="sourceLineNo">036</span>@InterfaceAudience.Public<a name="line.36"></a>
<span class="sourceLineNo">037</span>public class GroupingTableMap extends MapReduceBase<a name="line.37"></a>
<span class="sourceLineNo">038</span> implements TableMap&lt;ImmutableBytesWritable, Result&gt; {<a name="line.38"></a>
<span class="sourceLineNo">039</span><a name="line.39"></a>
<span class="sourceLineNo">040</span> /**<a name="line.40"></a>
<span class="sourceLineNo">041</span> * JobConf parameter to specify the columns used to produce the key passed to collect from the map<a name="line.41"></a>
<span class="sourceLineNo">042</span> * phase<a name="line.42"></a>
<span class="sourceLineNo">043</span> */<a name="line.43"></a>
<span class="sourceLineNo">044</span> public static final String GROUP_COLUMNS = "hbase.mapred.groupingtablemap.columns";<a name="line.44"></a>
<span class="sourceLineNo">045</span><a name="line.45"></a>
<span class="sourceLineNo">046</span> protected byte[][] columns;<a name="line.46"></a>
<span class="sourceLineNo">047</span><a name="line.47"></a>
<span class="sourceLineNo">048</span> /**<a name="line.48"></a>
<span class="sourceLineNo">049</span> * Use this before submitting a TableMap job. It will appropriately set up the JobConf.<a name="line.49"></a>
<span class="sourceLineNo">050</span> * @param table table to be processed<a name="line.50"></a>
<span class="sourceLineNo">051</span> * @param columns space separated list of columns to fetch<a name="line.51"></a>
<span class="sourceLineNo">052</span> * @param groupColumns space separated list of columns used to form the key used in collect<a name="line.52"></a>
<span class="sourceLineNo">053</span> * @param mapper map class<a name="line.53"></a>
<span class="sourceLineNo">054</span> * @param job job configuration object<a name="line.54"></a>
<span class="sourceLineNo">055</span> */<a name="line.55"></a>
<span class="sourceLineNo">056</span> @SuppressWarnings("unchecked")<a name="line.56"></a>
<span class="sourceLineNo">057</span> public static void initJob(String table, String columns, String groupColumns,<a name="line.57"></a>
<span class="sourceLineNo">058</span> Class&lt;? extends TableMap&gt; mapper, JobConf job) {<a name="line.58"></a>
<span class="sourceLineNo">059</span><a name="line.59"></a>
<span class="sourceLineNo">060</span> TableMapReduceUtil.initTableMapJob(table, columns, mapper, ImmutableBytesWritable.class,<a name="line.60"></a>
<span class="sourceLineNo">061</span> Result.class, job);<a name="line.61"></a>
<span class="sourceLineNo">062</span> job.set(GROUP_COLUMNS, groupColumns);<a name="line.62"></a>
<span class="sourceLineNo">063</span> }<a name="line.63"></a>
<span class="sourceLineNo">064</span><a name="line.64"></a>
<span class="sourceLineNo">065</span> @Override<a name="line.65"></a>
<span class="sourceLineNo">066</span> public void configure(JobConf job) {<a name="line.66"></a>
<span class="sourceLineNo">067</span> super.configure(job);<a name="line.67"></a>
<span class="sourceLineNo">068</span> String[] cols = job.get(GROUP_COLUMNS, "").split(" ");<a name="line.68"></a>
<span class="sourceLineNo">069</span> columns = new byte[cols.length][];<a name="line.69"></a>
<span class="sourceLineNo">070</span> for (int i = 0; i &lt; cols.length; i++) {<a name="line.70"></a>
<span class="sourceLineNo">071</span> columns[i] = Bytes.toBytes(cols[i]);<a name="line.71"></a>
<span class="sourceLineNo">072</span> }<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> /**<a name="line.75"></a>
<span class="sourceLineNo">076</span> * Extract the grouping columns from value to construct a new key. Pass the new key and value to<a name="line.76"></a>
<span class="sourceLineNo">077</span> * reduce. If any of the grouping columns are not found in the value, the record is skipped.<a name="line.77"></a>
<span class="sourceLineNo">078</span> */<a name="line.78"></a>
<span class="sourceLineNo">079</span> public void map(ImmutableBytesWritable key, Result value,<a name="line.79"></a>
<span class="sourceLineNo">080</span> OutputCollector&lt;ImmutableBytesWritable, Result&gt; output, Reporter reporter) throws IOException {<a name="line.80"></a>
<span class="sourceLineNo">081</span><a name="line.81"></a>
<span class="sourceLineNo">082</span> byte[][] keyVals = extractKeyValues(value);<a name="line.82"></a>
<span class="sourceLineNo">083</span> if (keyVals != null) {<a name="line.83"></a>
<span class="sourceLineNo">084</span> ImmutableBytesWritable tKey = createGroupKey(keyVals);<a name="line.84"></a>
<span class="sourceLineNo">085</span> output.collect(tKey, value);<a name="line.85"></a>
<span class="sourceLineNo">086</span> }<a name="line.86"></a>
<span class="sourceLineNo">087</span> }<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> * Extract columns values from the current record. This method returns null if any of the columns<a name="line.90"></a>
<span class="sourceLineNo">091</span> * are not found. Override this method if you want to deal with nulls differently.<a name="line.91"></a>
<span class="sourceLineNo">092</span> * @return array of byte values<a name="line.92"></a>
<span class="sourceLineNo">093</span> */<a name="line.93"></a>
<span class="sourceLineNo">094</span> protected byte[][] extractKeyValues(Result r) {<a name="line.94"></a>
<span class="sourceLineNo">095</span> byte[][] keyVals = null;<a name="line.95"></a>
<span class="sourceLineNo">096</span> ArrayList&lt;byte[]&gt; foundList = new ArrayList&lt;&gt;();<a name="line.96"></a>
<span class="sourceLineNo">097</span> int numCols = columns.length;<a name="line.97"></a>
<span class="sourceLineNo">098</span> if (numCols &gt; 0) {<a name="line.98"></a>
<span class="sourceLineNo">099</span> for (Cell value : r.listCells()) {<a name="line.99"></a>
<span class="sourceLineNo">100</span> byte[] column =<a name="line.100"></a>
<span class="sourceLineNo">101</span> CellUtil.makeColumn(CellUtil.cloneFamily(value), CellUtil.cloneQualifier(value));<a name="line.101"></a>
<span class="sourceLineNo">102</span> for (int i = 0; i &lt; numCols; i++) {<a name="line.102"></a>
<span class="sourceLineNo">103</span> if (Bytes.equals(column, columns[i])) {<a name="line.103"></a>
<span class="sourceLineNo">104</span> foundList.add(CellUtil.cloneValue(value));<a name="line.104"></a>
<span class="sourceLineNo">105</span> break;<a name="line.105"></a>
<span class="sourceLineNo">106</span> }<a name="line.106"></a>
<span class="sourceLineNo">107</span> }<a name="line.107"></a>
<span class="sourceLineNo">108</span> }<a name="line.108"></a>
<span class="sourceLineNo">109</span> if (foundList.size() == numCols) {<a name="line.109"></a>
<span class="sourceLineNo">110</span> keyVals = foundList.toArray(new byte[numCols][]);<a name="line.110"></a>
<span class="sourceLineNo">111</span> }<a name="line.111"></a>
<span class="sourceLineNo">112</span> }<a name="line.112"></a>
<span class="sourceLineNo">113</span> return keyVals;<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> /**<a name="line.116"></a>
<span class="sourceLineNo">117</span> * Create a key by concatenating multiple column values. Override this function in order to<a name="line.117"></a>
<span class="sourceLineNo">118</span> * produce different types of keys.<a name="line.118"></a>
<span class="sourceLineNo">119</span> * @return key generated by concatenating multiple column values<a name="line.119"></a>
<span class="sourceLineNo">120</span> */<a name="line.120"></a>
<span class="sourceLineNo">121</span> protected ImmutableBytesWritable createGroupKey(byte[][] vals) {<a name="line.121"></a>
<span class="sourceLineNo">122</span> if (vals == null) {<a name="line.122"></a>
<span class="sourceLineNo">123</span> return null;<a name="line.123"></a>
<span class="sourceLineNo">124</span> }<a name="line.124"></a>
<span class="sourceLineNo">125</span> StringBuilder sb = new StringBuilder();<a name="line.125"></a>
<span class="sourceLineNo">126</span> for (int i = 0; i &lt; vals.length; i++) {<a name="line.126"></a>
<span class="sourceLineNo">127</span> if (i &gt; 0) {<a name="line.127"></a>
<span class="sourceLineNo">128</span> sb.append(" ");<a name="line.128"></a>
<span class="sourceLineNo">129</span> }<a name="line.129"></a>
<span class="sourceLineNo">130</span> sb.append(Bytes.toString(vals[i]));<a name="line.130"></a>
<span class="sourceLineNo">131</span> }<a name="line.131"></a>
<span class="sourceLineNo">132</span> return new ImmutableBytesWritable(Bytes.toBytesBinary(sb.toString()));<a name="line.132"></a>
<span class="sourceLineNo">133</span> }<a name="line.133"></a>
<span class="sourceLineNo">134</span>}<a name="line.134"></a>
</pre>
</div>
</body>
</html>