blob: f2442888cee8cf987c3cf15214adbf9c28a0c80a [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.nio.charset.StandardCharsets;<a name="line.21"></a>
<span class="sourceLineNo">022</span>import org.apache.hadoop.hbase.KeyValue;<a name="line.22"></a>
<span class="sourceLineNo">023</span>import org.apache.hadoop.hbase.client.Durability;<a name="line.23"></a>
<span class="sourceLineNo">024</span>import org.apache.hadoop.hbase.client.Put;<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.io.LongWritable;<a name="line.27"></a>
<span class="sourceLineNo">028</span>import org.apache.hadoop.io.Text;<a name="line.28"></a>
<span class="sourceLineNo">029</span><a name="line.29"></a>
<span class="sourceLineNo">030</span>/**<a name="line.30"></a>
<span class="sourceLineNo">031</span> * Dummy mapper used for unit tests to verify that the mapper can be injected. This approach would<a name="line.31"></a>
<span class="sourceLineNo">032</span> * be used if a custom transformation needed to be done after reading the input data before writing<a name="line.32"></a>
<span class="sourceLineNo">033</span> * it to HFiles.<a name="line.33"></a>
<span class="sourceLineNo">034</span> */<a name="line.34"></a>
<span class="sourceLineNo">035</span>public class TsvImporterCustomTestMapper extends TsvImporterMapper {<a name="line.35"></a>
<span class="sourceLineNo">036</span> @Override<a name="line.36"></a>
<span class="sourceLineNo">037</span> protected void setup(Context context) {<a name="line.37"></a>
<span class="sourceLineNo">038</span> doSetup(context);<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> /**<a name="line.41"></a>
<span class="sourceLineNo">042</span> * Convert a line of TSV text into an HBase table row after transforming the values by multiplying<a name="line.42"></a>
<span class="sourceLineNo">043</span> * them by 3.<a name="line.43"></a>
<span class="sourceLineNo">044</span> */<a name="line.44"></a>
<span class="sourceLineNo">045</span> @Override<a name="line.45"></a>
<span class="sourceLineNo">046</span> public void map(LongWritable offset, Text value, Context context) throws IOException {<a name="line.46"></a>
<span class="sourceLineNo">047</span> byte[] family = Bytes.toBytes("FAM");<a name="line.47"></a>
<span class="sourceLineNo">048</span> final byte[][] qualifiers = { Bytes.toBytes("A"), Bytes.toBytes("B") };<a name="line.48"></a>
<span class="sourceLineNo">049</span><a name="line.49"></a>
<span class="sourceLineNo">050</span> // do some basic line parsing<a name="line.50"></a>
<span class="sourceLineNo">051</span> byte[] lineBytes = value.getBytes();<a name="line.51"></a>
<span class="sourceLineNo">052</span> String[] valueTokens = new String(lineBytes, StandardCharsets.UTF_8).split("\u001b");<a name="line.52"></a>
<span class="sourceLineNo">053</span><a name="line.53"></a>
<span class="sourceLineNo">054</span> // create the rowKey and Put<a name="line.54"></a>
<span class="sourceLineNo">055</span> ImmutableBytesWritable rowKey = new ImmutableBytesWritable(Bytes.toBytes(valueTokens[0]));<a name="line.55"></a>
<span class="sourceLineNo">056</span> Put put = new Put(rowKey.copyBytes());<a name="line.56"></a>
<span class="sourceLineNo">057</span> put.setDurability(Durability.SKIP_WAL);<a name="line.57"></a>
<span class="sourceLineNo">058</span><a name="line.58"></a>
<span class="sourceLineNo">059</span> // The value should look like this: VALUE1 or VALUE2. Let's multiply<a name="line.59"></a>
<span class="sourceLineNo">060</span> // the integer by 3<a name="line.60"></a>
<span class="sourceLineNo">061</span> for (int i = 1; i &lt; valueTokens.length; i++) {<a name="line.61"></a>
<span class="sourceLineNo">062</span> String prefix = valueTokens[i].substring(0, "VALUE".length());<a name="line.62"></a>
<span class="sourceLineNo">063</span> String suffix = valueTokens[i].substring("VALUE".length());<a name="line.63"></a>
<span class="sourceLineNo">064</span> String newValue = prefix + Integer.parseInt(suffix) * 3;<a name="line.64"></a>
<span class="sourceLineNo">065</span><a name="line.65"></a>
<span class="sourceLineNo">066</span> KeyValue kv =<a name="line.66"></a>
<span class="sourceLineNo">067</span> new KeyValue(rowKey.copyBytes(), family, qualifiers[i - 1], Bytes.toBytes(newValue));<a name="line.67"></a>
<span class="sourceLineNo">068</span> put.add(kv);<a name="line.68"></a>
<span class="sourceLineNo">069</span> }<a name="line.69"></a>
<span class="sourceLineNo">070</span><a name="line.70"></a>
<span class="sourceLineNo">071</span> try {<a name="line.71"></a>
<span class="sourceLineNo">072</span> context.write(rowKey, put);<a name="line.72"></a>
<span class="sourceLineNo">073</span> } catch (InterruptedException e) {<a name="line.73"></a>
<span class="sourceLineNo">074</span> e.printStackTrace();<a name="line.74"></a>
<span class="sourceLineNo">075</span> }<a name="line.75"></a>
<span class="sourceLineNo">076</span> }<a name="line.76"></a>
<span class="sourceLineNo">077</span>}<a name="line.77"></a>
</pre>
</div>
</body>
</html>