blob: 15fa12a36ca1c2e7c9061c09dc4c654014149056 [file] [log] [blame]
<!DOCTYPE HTML>
<html lang="en">
<head>
<!-- Generated by javadoc (17) -->
<title>Source code</title>
<meta name="viewport" content="width=device-width, initial-scale=1">
<meta name="description" content="source: package: org.apache.hadoop.hbase.mapred, class: TableInputFormat">
<meta name="generator" content="javadoc/SourceToHTMLConverter">
<link rel="stylesheet" type="text/css" href="../../../../../../stylesheet.css" title="Style">
</head>
<body class="source-page">
<main role="main">
<div class="source-container">
<pre><span class="source-line-no">001</span><span id="line-1">/*</span>
<span class="source-line-no">002</span><span id="line-2"> * Licensed to the Apache Software Foundation (ASF) under one</span>
<span class="source-line-no">003</span><span id="line-3"> * or more contributor license agreements. See the NOTICE file</span>
<span class="source-line-no">004</span><span id="line-4"> * distributed with this work for additional information</span>
<span class="source-line-no">005</span><span id="line-5"> * regarding copyright ownership. The ASF licenses this file</span>
<span class="source-line-no">006</span><span id="line-6"> * to you under the Apache License, Version 2.0 (the</span>
<span class="source-line-no">007</span><span id="line-7"> * "License"); you may not use this file except in compliance</span>
<span class="source-line-no">008</span><span id="line-8"> * with the License. You may obtain a copy of the License at</span>
<span class="source-line-no">009</span><span id="line-9"> *</span>
<span class="source-line-no">010</span><span id="line-10"> * http://www.apache.org/licenses/LICENSE-2.0</span>
<span class="source-line-no">011</span><span id="line-11"> *</span>
<span class="source-line-no">012</span><span id="line-12"> * Unless required by applicable law or agreed to in writing, software</span>
<span class="source-line-no">013</span><span id="line-13"> * distributed under the License is distributed on an "AS IS" BASIS,</span>
<span class="source-line-no">014</span><span id="line-14"> * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.</span>
<span class="source-line-no">015</span><span id="line-15"> * See the License for the specific language governing permissions and</span>
<span class="source-line-no">016</span><span id="line-16"> * limitations under the License.</span>
<span class="source-line-no">017</span><span id="line-17"> */</span>
<span class="source-line-no">018</span><span id="line-18">package org.apache.hadoop.hbase.mapred;</span>
<span class="source-line-no">019</span><span id="line-19"></span>
<span class="source-line-no">020</span><span id="line-20">import java.io.IOException;</span>
<span class="source-line-no">021</span><span id="line-21">import org.apache.hadoop.fs.Path;</span>
<span class="source-line-no">022</span><span id="line-22">import org.apache.hadoop.hbase.TableName;</span>
<span class="source-line-no">023</span><span id="line-23">import org.apache.hadoop.hbase.client.Connection;</span>
<span class="source-line-no">024</span><span id="line-24">import org.apache.hadoop.hbase.client.ConnectionFactory;</span>
<span class="source-line-no">025</span><span id="line-25">import org.apache.hadoop.hbase.util.Bytes;</span>
<span class="source-line-no">026</span><span id="line-26">import org.apache.hadoop.mapred.FileInputFormat;</span>
<span class="source-line-no">027</span><span id="line-27">import org.apache.hadoop.mapred.JobConf;</span>
<span class="source-line-no">028</span><span id="line-28">import org.apache.hadoop.mapred.JobConfigurable;</span>
<span class="source-line-no">029</span><span id="line-29">import org.apache.hadoop.util.StringUtils;</span>
<span class="source-line-no">030</span><span id="line-30">import org.apache.yetus.audience.InterfaceAudience;</span>
<span class="source-line-no">031</span><span id="line-31">import org.slf4j.Logger;</span>
<span class="source-line-no">032</span><span id="line-32">import org.slf4j.LoggerFactory;</span>
<span class="source-line-no">033</span><span id="line-33"></span>
<span class="source-line-no">034</span><span id="line-34">/**</span>
<span class="source-line-no">035</span><span id="line-35"> * Convert HBase tabular data into a format that is consumable by Map/Reduce.</span>
<span class="source-line-no">036</span><span id="line-36"> */</span>
<span class="source-line-no">037</span><span id="line-37">@InterfaceAudience.Public</span>
<span class="source-line-no">038</span><span id="line-38">public class TableInputFormat extends TableInputFormatBase implements JobConfigurable {</span>
<span class="source-line-no">039</span><span id="line-39"> private static final Logger LOG = LoggerFactory.getLogger(TableInputFormat.class);</span>
<span class="source-line-no">040</span><span id="line-40"></span>
<span class="source-line-no">041</span><span id="line-41"> /**</span>
<span class="source-line-no">042</span><span id="line-42"> * space delimited list of columns</span>
<span class="source-line-no">043</span><span id="line-43"> */</span>
<span class="source-line-no">044</span><span id="line-44"> public static final String COLUMN_LIST = "hbase.mapred.tablecolumns";</span>
<span class="source-line-no">045</span><span id="line-45"></span>
<span class="source-line-no">046</span><span id="line-46"> public void configure(JobConf job) {</span>
<span class="source-line-no">047</span><span id="line-47"> try {</span>
<span class="source-line-no">048</span><span id="line-48"> initialize(job);</span>
<span class="source-line-no">049</span><span id="line-49"> } catch (Exception e) {</span>
<span class="source-line-no">050</span><span id="line-50"> LOG.error(StringUtils.stringifyException(e));</span>
<span class="source-line-no">051</span><span id="line-51"> }</span>
<span class="source-line-no">052</span><span id="line-52"> }</span>
<span class="source-line-no">053</span><span id="line-53"></span>
<span class="source-line-no">054</span><span id="line-54"> @Override</span>
<span class="source-line-no">055</span><span id="line-55"> protected void initialize(JobConf job) throws IOException {</span>
<span class="source-line-no">056</span><span id="line-56"> Path[] tableNames = FileInputFormat.getInputPaths(job);</span>
<span class="source-line-no">057</span><span id="line-57"> String colArg = job.get(COLUMN_LIST);</span>
<span class="source-line-no">058</span><span id="line-58"> String[] colNames = colArg.split(" ");</span>
<span class="source-line-no">059</span><span id="line-59"> byte[][] m_cols = new byte[colNames.length][];</span>
<span class="source-line-no">060</span><span id="line-60"> for (int i = 0; i &lt; m_cols.length; i++) {</span>
<span class="source-line-no">061</span><span id="line-61"> m_cols[i] = Bytes.toBytes(colNames[i]);</span>
<span class="source-line-no">062</span><span id="line-62"> }</span>
<span class="source-line-no">063</span><span id="line-63"> setInputColumns(m_cols);</span>
<span class="source-line-no">064</span><span id="line-64"> Connection connection = ConnectionFactory.createConnection(job);</span>
<span class="source-line-no">065</span><span id="line-65"> initializeTable(connection, TableName.valueOf(tableNames[0].getName()));</span>
<span class="source-line-no">066</span><span id="line-66"> }</span>
<span class="source-line-no">067</span><span id="line-67"></span>
<span class="source-line-no">068</span><span id="line-68"> public void validateInput(JobConf job) throws IOException {</span>
<span class="source-line-no">069</span><span id="line-69"> // expecting exactly one path</span>
<span class="source-line-no">070</span><span id="line-70"> Path[] tableNames = FileInputFormat.getInputPaths(job);</span>
<span class="source-line-no">071</span><span id="line-71"> if (tableNames == null || tableNames.length &gt; 1) {</span>
<span class="source-line-no">072</span><span id="line-72"> throw new IOException("expecting one table name");</span>
<span class="source-line-no">073</span><span id="line-73"> }</span>
<span class="source-line-no">074</span><span id="line-74"></span>
<span class="source-line-no">075</span><span id="line-75"> // connected to table?</span>
<span class="source-line-no">076</span><span id="line-76"> if (getTable() == null) {</span>
<span class="source-line-no">077</span><span id="line-77"> throw new IOException("could not connect to table '" + tableNames[0].getName() + "'");</span>
<span class="source-line-no">078</span><span id="line-78"> }</span>
<span class="source-line-no">079</span><span id="line-79"></span>
<span class="source-line-no">080</span><span id="line-80"> // expecting at least one column</span>
<span class="source-line-no">081</span><span id="line-81"> String colArg = job.get(COLUMN_LIST);</span>
<span class="source-line-no">082</span><span id="line-82"> if (colArg == null || colArg.length() == 0) {</span>
<span class="source-line-no">083</span><span id="line-83"> throw new IOException("expecting at least one column");</span>
<span class="source-line-no">084</span><span id="line-84"> }</span>
<span class="source-line-no">085</span><span id="line-85"> }</span>
<span class="source-line-no">086</span><span id="line-86">}</span>
</pre>
</div>
</main>
</body>
</html>