blob: a7456f6d950de6e7da09bbb287341eb2903996cd [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.mapreduce;<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 java.util.List;<a name="line.22"></a>
<span class="sourceLineNo">023</span>import org.apache.hadoop.conf.Configurable;<a name="line.23"></a>
<span class="sourceLineNo">024</span>import org.apache.hadoop.conf.Configuration;<a name="line.24"></a>
<span class="sourceLineNo">025</span>import org.apache.hadoop.hbase.client.Scan;<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><a name="line.27"></a>
<span class="sourceLineNo">028</span>/**<a name="line.28"></a>
<span class="sourceLineNo">029</span> * Convert HBase tabular data from multiple scanners into a format that is consumable by Map/Reduce.<a name="line.29"></a>
<span class="sourceLineNo">030</span> * &lt;p&gt;<a name="line.30"></a>
<span class="sourceLineNo">031</span> * Usage example<a name="line.31"></a>
<span class="sourceLineNo">032</span> * &lt;/p&gt;<a name="line.32"></a>
<span class="sourceLineNo">033</span> *<a name="line.33"></a>
<span class="sourceLineNo">034</span> * &lt;pre&gt;<a name="line.34"></a>
<span class="sourceLineNo">035</span> * List&amp;lt;Scan&amp;gt; scans = new ArrayList&amp;lt;Scan&amp;gt;();<a name="line.35"></a>
<span class="sourceLineNo">036</span> *<a name="line.36"></a>
<span class="sourceLineNo">037</span> * Scan scan1 = new Scan();<a name="line.37"></a>
<span class="sourceLineNo">038</span> * scan1.withStartRow(firstRow1);<a name="line.38"></a>
<span class="sourceLineNo">039</span> * scan1.withStopRow(lastRow1);<a name="line.39"></a>
<span class="sourceLineNo">040</span> * scan1.setAttribute(Scan.SCAN_ATTRIBUTES_TABLE_NAME, table1);<a name="line.40"></a>
<span class="sourceLineNo">041</span> * scans.add(scan1);<a name="line.41"></a>
<span class="sourceLineNo">042</span> *<a name="line.42"></a>
<span class="sourceLineNo">043</span> * Scan scan2 = new Scan();<a name="line.43"></a>
<span class="sourceLineNo">044</span> * scan2.withStartRow(firstRow2);<a name="line.44"></a>
<span class="sourceLineNo">045</span> * scan2.withStopRow(lastRow2);<a name="line.45"></a>
<span class="sourceLineNo">046</span> * scan1.setAttribute(Scan.SCAN_ATTRIBUTES_TABLE_NAME, table2);<a name="line.46"></a>
<span class="sourceLineNo">047</span> * scans.add(scan2);<a name="line.47"></a>
<span class="sourceLineNo">048</span> *<a name="line.48"></a>
<span class="sourceLineNo">049</span> * TableMapReduceUtil.initTableMapperJob(scans, TableMapper.class, Text.class, IntWritable.class,<a name="line.49"></a>
<span class="sourceLineNo">050</span> * job);<a name="line.50"></a>
<span class="sourceLineNo">051</span> * &lt;/pre&gt;<a name="line.51"></a>
<span class="sourceLineNo">052</span> */<a name="line.52"></a>
<span class="sourceLineNo">053</span>@InterfaceAudience.Public<a name="line.53"></a>
<span class="sourceLineNo">054</span>public class MultiTableInputFormat extends MultiTableInputFormatBase implements Configurable {<a name="line.54"></a>
<span class="sourceLineNo">055</span><a name="line.55"></a>
<span class="sourceLineNo">056</span> /** Job parameter that specifies the scan list. */<a name="line.56"></a>
<span class="sourceLineNo">057</span> public static final String SCANS = "hbase.mapreduce.scans";<a name="line.57"></a>
<span class="sourceLineNo">058</span><a name="line.58"></a>
<span class="sourceLineNo">059</span> /** The configuration. */<a name="line.59"></a>
<span class="sourceLineNo">060</span> private Configuration conf = null;<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> * Returns the current configuration.<a name="line.63"></a>
<span class="sourceLineNo">064</span> * @return The current configuration.<a name="line.64"></a>
<span class="sourceLineNo">065</span> * @see org.apache.hadoop.conf.Configurable#getConf()<a name="line.65"></a>
<span class="sourceLineNo">066</span> */<a name="line.66"></a>
<span class="sourceLineNo">067</span> @Override<a name="line.67"></a>
<span class="sourceLineNo">068</span> public Configuration getConf() {<a name="line.68"></a>
<span class="sourceLineNo">069</span> return conf;<a name="line.69"></a>
<span class="sourceLineNo">070</span> }<a name="line.70"></a>
<span class="sourceLineNo">071</span><a name="line.71"></a>
<span class="sourceLineNo">072</span> /**<a name="line.72"></a>
<span class="sourceLineNo">073</span> * Sets the configuration. This is used to set the details for the tables to be scanned.<a name="line.73"></a>
<span class="sourceLineNo">074</span> * @param configuration The configuration to set.<a name="line.74"></a>
<span class="sourceLineNo">075</span> * @see org.apache.hadoop.conf.Configurable#setConf( org.apache.hadoop.conf.Configuration)<a name="line.75"></a>
<span class="sourceLineNo">076</span> */<a name="line.76"></a>
<span class="sourceLineNo">077</span> @Override<a name="line.77"></a>
<span class="sourceLineNo">078</span> public void setConf(Configuration configuration) {<a name="line.78"></a>
<span class="sourceLineNo">079</span> this.conf = configuration;<a name="line.79"></a>
<span class="sourceLineNo">080</span> String[] rawScans = conf.getStrings(SCANS);<a name="line.80"></a>
<span class="sourceLineNo">081</span> if (rawScans.length &lt;= 0) {<a name="line.81"></a>
<span class="sourceLineNo">082</span> throw new IllegalArgumentException(<a name="line.82"></a>
<span class="sourceLineNo">083</span> "There must be at least 1 scan configuration set to : " + SCANS);<a name="line.83"></a>
<span class="sourceLineNo">084</span> }<a name="line.84"></a>
<span class="sourceLineNo">085</span> List&lt;Scan&gt; scans = new ArrayList&lt;&gt;();<a name="line.85"></a>
<span class="sourceLineNo">086</span><a name="line.86"></a>
<span class="sourceLineNo">087</span> for (int i = 0; i &lt; rawScans.length; i++) {<a name="line.87"></a>
<span class="sourceLineNo">088</span> try {<a name="line.88"></a>
<span class="sourceLineNo">089</span> scans.add(TableMapReduceUtil.convertStringToScan(rawScans[i]));<a name="line.89"></a>
<span class="sourceLineNo">090</span> } catch (IOException e) {<a name="line.90"></a>
<span class="sourceLineNo">091</span> throw new RuntimeException("Failed to convert Scan : " + rawScans[i] + " to string", e);<a name="line.91"></a>
<span class="sourceLineNo">092</span> }<a name="line.92"></a>
<span class="sourceLineNo">093</span> }<a name="line.93"></a>
<span class="sourceLineNo">094</span> this.setScans(scans);<a name="line.94"></a>
<span class="sourceLineNo">095</span> }<a name="line.95"></a>
<span class="sourceLineNo">096</span>}<a name="line.96"></a>
</pre>
</div>
</body>
</html>