blob: 5c1fd4841a1fe153b04ab45f3b0c15b77210a916 [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 org.apache.commons.lang3.ArrayUtils;<a name="line.21"></a>
<span class="sourceLineNo">022</span>import org.apache.hadoop.conf.Configuration;<a name="line.22"></a>
<span class="sourceLineNo">023</span>import org.apache.hadoop.conf.Configured;<a name="line.23"></a>
<span class="sourceLineNo">024</span>import org.apache.hadoop.fs.Path;<a name="line.24"></a>
<span class="sourceLineNo">025</span>import org.apache.hadoop.hbase.HBaseConfiguration;<a name="line.25"></a>
<span class="sourceLineNo">026</span>import org.apache.hadoop.hbase.TableName;<a name="line.26"></a>
<span class="sourceLineNo">027</span>import org.apache.hadoop.hbase.client.Result;<a name="line.27"></a>
<span class="sourceLineNo">028</span>import org.apache.hadoop.hbase.client.Scan;<a name="line.28"></a>
<span class="sourceLineNo">029</span>import org.apache.hadoop.hbase.io.ImmutableBytesWritable;<a name="line.29"></a>
<span class="sourceLineNo">030</span>import org.apache.hadoop.hbase.util.Triple;<a name="line.30"></a>
<span class="sourceLineNo">031</span>import org.apache.hadoop.mapreduce.Job;<a name="line.31"></a>
<span class="sourceLineNo">032</span>import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat;<a name="line.32"></a>
<span class="sourceLineNo">033</span>import org.apache.hadoop.mapreduce.lib.output.SequenceFileOutputFormat;<a name="line.33"></a>
<span class="sourceLineNo">034</span>import org.apache.hadoop.util.Tool;<a name="line.34"></a>
<span class="sourceLineNo">035</span>import org.apache.hadoop.util.ToolRunner;<a name="line.35"></a>
<span class="sourceLineNo">036</span>import org.apache.yetus.audience.InterfaceAudience;<a name="line.36"></a>
<span class="sourceLineNo">037</span><a name="line.37"></a>
<span class="sourceLineNo">038</span>/**<a name="line.38"></a>
<span class="sourceLineNo">039</span> * Export an HBase table. Writes content to sequence files up in HDFS. Use {@link Import} to read it<a name="line.39"></a>
<span class="sourceLineNo">040</span> * back in again.<a name="line.40"></a>
<span class="sourceLineNo">041</span> */<a name="line.41"></a>
<span class="sourceLineNo">042</span>@InterfaceAudience.Public<a name="line.42"></a>
<span class="sourceLineNo">043</span>public class Export extends Configured implements Tool {<a name="line.43"></a>
<span class="sourceLineNo">044</span> static final String NAME = "export";<a name="line.44"></a>
<span class="sourceLineNo">045</span> static final String JOB_NAME_CONF_KEY = "mapreduce.job.name";<a name="line.45"></a>
<span class="sourceLineNo">046</span><a name="line.46"></a>
<span class="sourceLineNo">047</span> /**<a name="line.47"></a>
<span class="sourceLineNo">048</span> * Sets up the actual job.<a name="line.48"></a>
<span class="sourceLineNo">049</span> * @param conf The current configuration.<a name="line.49"></a>
<span class="sourceLineNo">050</span> * @param args The command line parameters.<a name="line.50"></a>
<span class="sourceLineNo">051</span> * @return The newly created job.<a name="line.51"></a>
<span class="sourceLineNo">052</span> * @throws IOException When setting up the job fails.<a name="line.52"></a>
<span class="sourceLineNo">053</span> */<a name="line.53"></a>
<span class="sourceLineNo">054</span> public static Job createSubmittableJob(Configuration conf, String[] args) throws IOException {<a name="line.54"></a>
<span class="sourceLineNo">055</span> Triple&lt;TableName, Scan, Path&gt; arguments = ExportUtils.getArgumentsFromCommandLine(conf, args);<a name="line.55"></a>
<span class="sourceLineNo">056</span> String tableName = arguments.getFirst().getNameAsString();<a name="line.56"></a>
<span class="sourceLineNo">057</span> Path outputDir = arguments.getThird();<a name="line.57"></a>
<span class="sourceLineNo">058</span> Job job = Job.getInstance(conf, conf.get(JOB_NAME_CONF_KEY, NAME + "_" + tableName));<a name="line.58"></a>
<span class="sourceLineNo">059</span> job.setJobName(NAME + "_" + tableName);<a name="line.59"></a>
<span class="sourceLineNo">060</span> job.setJarByClass(Export.class);<a name="line.60"></a>
<span class="sourceLineNo">061</span> // Set optional scan parameters<a name="line.61"></a>
<span class="sourceLineNo">062</span> Scan s = arguments.getSecond();<a name="line.62"></a>
<span class="sourceLineNo">063</span> IdentityTableMapper.initJob(tableName, s, IdentityTableMapper.class, job);<a name="line.63"></a>
<span class="sourceLineNo">064</span> // No reducers. Just write straight to output files.<a name="line.64"></a>
<span class="sourceLineNo">065</span> job.setNumReduceTasks(0);<a name="line.65"></a>
<span class="sourceLineNo">066</span> job.setOutputFormatClass(SequenceFileOutputFormat.class);<a name="line.66"></a>
<span class="sourceLineNo">067</span> job.setOutputKeyClass(ImmutableBytesWritable.class);<a name="line.67"></a>
<span class="sourceLineNo">068</span> job.setOutputValueClass(Result.class);<a name="line.68"></a>
<span class="sourceLineNo">069</span> FileOutputFormat.setOutputPath(job, outputDir); // job conf doesn't contain the conf so doesn't<a name="line.69"></a>
<span class="sourceLineNo">070</span> // have a default fs.<a name="line.70"></a>
<span class="sourceLineNo">071</span> return job;<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> @Override<a name="line.74"></a>
<span class="sourceLineNo">075</span> public int run(String[] args) throws Exception {<a name="line.75"></a>
<span class="sourceLineNo">076</span> if (!ExportUtils.isValidArguements(args)) {<a name="line.76"></a>
<span class="sourceLineNo">077</span> ExportUtils.usage("Wrong number of arguments: " + ArrayUtils.getLength(args));<a name="line.77"></a>
<span class="sourceLineNo">078</span> System.err.println(" -D " + JOB_NAME_CONF_KEY<a name="line.78"></a>
<span class="sourceLineNo">079</span> + "=jobName - use the specified mapreduce job name for the export");<a name="line.79"></a>
<span class="sourceLineNo">080</span> System.err.println("For MR performance consider the following properties:");<a name="line.80"></a>
<span class="sourceLineNo">081</span> System.err.println(" -D mapreduce.map.speculative=false");<a name="line.81"></a>
<span class="sourceLineNo">082</span> System.err.println(" -D mapreduce.reduce.speculative=false");<a name="line.82"></a>
<span class="sourceLineNo">083</span> return -1;<a name="line.83"></a>
<span class="sourceLineNo">084</span> }<a name="line.84"></a>
<span class="sourceLineNo">085</span> Job job = createSubmittableJob(getConf(), args);<a name="line.85"></a>
<span class="sourceLineNo">086</span> return (job.waitForCompletion(true) ? 0 : 1);<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> * Main entry point.<a name="line.90"></a>
<span class="sourceLineNo">091</span> * @param args The command line parameters.<a name="line.91"></a>
<span class="sourceLineNo">092</span> * @throws Exception When running the job fails.<a name="line.92"></a>
<span class="sourceLineNo">093</span> */<a name="line.93"></a>
<span class="sourceLineNo">094</span> public static void main(String[] args) throws Exception {<a name="line.94"></a>
<span class="sourceLineNo">095</span> int errCode = ToolRunner.run(HBaseConfiguration.create(), new Export(), args);<a name="line.95"></a>
<span class="sourceLineNo">096</span> System.exit(errCode);<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>