| <!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: MultiTableSnapshotInputFormat"> |
| <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 edu.umd.cs.findbugs.annotations.SuppressWarnings;</span> |
| <span class="source-line-no">021</span><span id="line-21">import java.io.IOException;</span> |
| <span class="source-line-no">022</span><span id="line-22">import java.util.Collection;</span> |
| <span class="source-line-no">023</span><span id="line-23">import java.util.List;</span> |
| <span class="source-line-no">024</span><span id="line-24">import java.util.Map;</span> |
| <span class="source-line-no">025</span><span id="line-25">import org.apache.hadoop.conf.Configuration;</span> |
| <span class="source-line-no">026</span><span id="line-26">import org.apache.hadoop.fs.Path;</span> |
| <span class="source-line-no">027</span><span id="line-27">import org.apache.hadoop.hbase.client.Result;</span> |
| <span class="source-line-no">028</span><span id="line-28">import org.apache.hadoop.hbase.client.Scan;</span> |
| <span class="source-line-no">029</span><span id="line-29">import org.apache.hadoop.hbase.io.ImmutableBytesWritable;</span> |
| <span class="source-line-no">030</span><span id="line-30">import org.apache.hadoop.hbase.mapreduce.MultiTableSnapshotInputFormatImpl;</span> |
| <span class="source-line-no">031</span><span id="line-31">import org.apache.hadoop.hbase.mapreduce.TableSnapshotInputFormatImpl;</span> |
| <span class="source-line-no">032</span><span id="line-32">import org.apache.hadoop.mapred.InputFormat;</span> |
| <span class="source-line-no">033</span><span id="line-33">import org.apache.hadoop.mapred.InputSplit;</span> |
| <span class="source-line-no">034</span><span id="line-34">import org.apache.hadoop.mapred.JobConf;</span> |
| <span class="source-line-no">035</span><span id="line-35">import org.apache.hadoop.mapred.RecordReader;</span> |
| <span class="source-line-no">036</span><span id="line-36">import org.apache.hadoop.mapred.Reporter;</span> |
| <span class="source-line-no">037</span><span id="line-37">import org.apache.yetus.audience.InterfaceAudience;</span> |
| <span class="source-line-no">038</span><span id="line-38"></span> |
| <span class="source-line-no">039</span><span id="line-39">/**</span> |
| <span class="source-line-no">040</span><span id="line-40"> * MultiTableSnapshotInputFormat generalizes</span> |
| <span class="source-line-no">041</span><span id="line-41"> * {@link org.apache.hadoop.hbase.mapred.TableSnapshotInputFormat} allowing a MapReduce job to run</span> |
| <span class="source-line-no">042</span><span id="line-42"> * over one or more table snapshots, with one or more scans configured for each. Internally, the</span> |
| <span class="source-line-no">043</span><span id="line-43"> * input format delegates to {@link org.apache.hadoop.hbase.mapreduce.TableSnapshotInputFormat} and</span> |
| <span class="source-line-no">044</span><span id="line-44"> * thus has the same performance advantages; see</span> |
| <span class="source-line-no">045</span><span id="line-45"> * {@link org.apache.hadoop.hbase.mapreduce.TableSnapshotInputFormat} for more details. Usage is</span> |
| <span class="source-line-no">046</span><span id="line-46"> * similar to TableSnapshotInputFormat, with the following exception:</span> |
| <span class="source-line-no">047</span><span id="line-47"> * initMultiTableSnapshotMapperJob takes in a map from snapshot name to a collection of scans. For</span> |
| <span class="source-line-no">048</span><span id="line-48"> * each snapshot in the map, each corresponding scan will be applied; the overall dataset for the</span> |
| <span class="source-line-no">049</span><span id="line-49"> * job is defined by the concatenation of the regions and tables included in each snapshot/scan</span> |
| <span class="source-line-no">050</span><span id="line-50"> * pair.</span> |
| <span class="source-line-no">051</span><span id="line-51"> * {@link TableMapReduceUtil#initMultiTableSnapshotMapperJob(Map, Class, Class, Class, JobConf, boolean, Path)}</span> |
| <span class="source-line-no">052</span><span id="line-52"> * can be used to configure the job.</span> |
| <span class="source-line-no">053</span><span id="line-53"> *</span> |
| <span class="source-line-no">054</span><span id="line-54"> * <pre></span> |
| <span class="source-line-no">055</span><span id="line-55"> * {@code</span> |
| <span class="source-line-no">056</span><span id="line-56"> * Job job = new Job(conf);</span> |
| <span class="source-line-no">057</span><span id="line-57"> * Map<String, Collection<Scan>> snapshotScans = ImmutableMap.of(</span> |
| <span class="source-line-no">058</span><span id="line-58"> * "snapshot1", ImmutableList.of(new Scan(Bytes.toBytes("a"), Bytes.toBytes("b"))),</span> |
| <span class="source-line-no">059</span><span id="line-59"> * "snapshot2", ImmutableList.of(new Scan(Bytes.toBytes("1"), Bytes.toBytes("2")))</span> |
| <span class="source-line-no">060</span><span id="line-60"> * );</span> |
| <span class="source-line-no">061</span><span id="line-61"> * Path restoreDir = new Path("/tmp/snapshot_restore_dir")</span> |
| <span class="source-line-no">062</span><span id="line-62"> * TableMapReduceUtil.initTableSnapshotMapperJob(</span> |
| <span class="source-line-no">063</span><span id="line-63"> * snapshotScans, MyTableMapper.class, MyMapKeyOutput.class,</span> |
| <span class="source-line-no">064</span><span id="line-64"> * MyMapOutputValueWritable.class, job, true, restoreDir);</span> |
| <span class="source-line-no">065</span><span id="line-65"> * }</span> |
| <span class="source-line-no">066</span><span id="line-66"> * </pre></span> |
| <span class="source-line-no">067</span><span id="line-67"> *</span> |
| <span class="source-line-no">068</span><span id="line-68"> * Internally, this input format restores each snapshot into a subdirectory of the given tmp</span> |
| <span class="source-line-no">069</span><span id="line-69"> * directory. Input splits and record readers are created as described in</span> |
| <span class="source-line-no">070</span><span id="line-70"> * {@link org.apache.hadoop.hbase.mapreduce.TableSnapshotInputFormat} (one per region). See</span> |
| <span class="source-line-no">071</span><span id="line-71"> * {@link org.apache.hadoop.hbase.mapreduce.TableSnapshotInputFormat} for more notes on</span> |
| <span class="source-line-no">072</span><span id="line-72"> * permissioning; the same caveats apply here.</span> |
| <span class="source-line-no">073</span><span id="line-73"> * @see org.apache.hadoop.hbase.mapreduce.TableSnapshotInputFormat</span> |
| <span class="source-line-no">074</span><span id="line-74"> * @see org.apache.hadoop.hbase.client.TableSnapshotScanner</span> |
| <span class="source-line-no">075</span><span id="line-75"> */</span> |
| <span class="source-line-no">076</span><span id="line-76">@InterfaceAudience.Public</span> |
| <span class="source-line-no">077</span><span id="line-77">public class MultiTableSnapshotInputFormat extends TableSnapshotInputFormat</span> |
| <span class="source-line-no">078</span><span id="line-78"> implements InputFormat<ImmutableBytesWritable, Result> {</span> |
| <span class="source-line-no">079</span><span id="line-79"></span> |
| <span class="source-line-no">080</span><span id="line-80"> private final MultiTableSnapshotInputFormatImpl delegate;</span> |
| <span class="source-line-no">081</span><span id="line-81"></span> |
| <span class="source-line-no">082</span><span id="line-82"> public MultiTableSnapshotInputFormat() {</span> |
| <span class="source-line-no">083</span><span id="line-83"> this.delegate = new MultiTableSnapshotInputFormatImpl();</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"> @Override</span> |
| <span class="source-line-no">087</span><span id="line-87"> public InputSplit[] getSplits(JobConf job, int numSplits) throws IOException {</span> |
| <span class="source-line-no">088</span><span id="line-88"> List<TableSnapshotInputFormatImpl.InputSplit> splits = delegate.getSplits(job);</span> |
| <span class="source-line-no">089</span><span id="line-89"> InputSplit[] results = new InputSplit[splits.size()];</span> |
| <span class="source-line-no">090</span><span id="line-90"> for (int i = 0; i < splits.size(); i++) {</span> |
| <span class="source-line-no">091</span><span id="line-91"> results[i] = new TableSnapshotRegionSplit(splits.get(i));</span> |
| <span class="source-line-no">092</span><span id="line-92"> }</span> |
| <span class="source-line-no">093</span><span id="line-93"> return results;</span> |
| <span class="source-line-no">094</span><span id="line-94"> }</span> |
| <span class="source-line-no">095</span><span id="line-95"></span> |
| <span class="source-line-no">096</span><span id="line-96"> @Override</span> |
| <span class="source-line-no">097</span><span id="line-97"> public RecordReader<ImmutableBytesWritable, Result> getRecordReader(InputSplit split, JobConf job,</span> |
| <span class="source-line-no">098</span><span id="line-98"> Reporter reporter) throws IOException {</span> |
| <span class="source-line-no">099</span><span id="line-99"> return new TableSnapshotRecordReader((TableSnapshotRegionSplit) split, job);</span> |
| <span class="source-line-no">100</span><span id="line-100"> }</span> |
| <span class="source-line-no">101</span><span id="line-101"></span> |
| <span class="source-line-no">102</span><span id="line-102"> @SuppressWarnings("checkstyle:linelength")</span> |
| <span class="source-line-no">103</span><span id="line-103"> /**</span> |
| <span class="source-line-no">104</span><span id="line-104"> * Configure conf to read from snapshotScans, with snapshots restored to a subdirectory of</span> |
| <span class="source-line-no">105</span><span id="line-105"> * restoreDir. Sets:</span> |
| <span class="source-line-no">106</span><span id="line-106"> * {@link org.apache.hadoop.hbase.mapreduce.MultiTableSnapshotInputFormatImpl#RESTORE_DIRS_KEY},</span> |
| <span class="source-line-no">107</span><span id="line-107"> * {@link org.apache.hadoop.hbase.mapreduce.MultiTableSnapshotInputFormatImpl#SNAPSHOT_TO_SCANS_KEY}</span> |
| <span class="source-line-no">108</span><span id="line-108"> */</span> |
| <span class="source-line-no">109</span><span id="line-109"> public static void setInput(Configuration conf, Map<String, Collection<Scan>> snapshotScans,</span> |
| <span class="source-line-no">110</span><span id="line-110"> Path restoreDir) throws IOException {</span> |
| <span class="source-line-no">111</span><span id="line-111"> new MultiTableSnapshotInputFormatImpl().setInput(conf, snapshotScans, restoreDir);</span> |
| <span class="source-line-no">112</span><span id="line-112"> }</span> |
| <span class="source-line-no">113</span><span id="line-113"></span> |
| <span class="source-line-no">114</span><span id="line-114">}</span> |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| </pre> |
| </div> |
| </main> |
| </body> |
| </html> |
