| <!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.mapreduce, class: MultiTableOutputFormat"> |
| <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.mapreduce;</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 java.util.HashMap;</span> |
| <span class="source-line-no">022</span><span id="line-22">import java.util.Map;</span> |
| <span class="source-line-no">023</span><span id="line-23">import org.apache.hadoop.conf.Configuration;</span> |
| <span class="source-line-no">024</span><span id="line-24">import org.apache.hadoop.hbase.HBaseConfiguration;</span> |
| <span class="source-line-no">025</span><span id="line-25">import org.apache.hadoop.hbase.TableName;</span> |
| <span class="source-line-no">026</span><span id="line-26">import org.apache.hadoop.hbase.client.BufferedMutator;</span> |
| <span class="source-line-no">027</span><span id="line-27">import org.apache.hadoop.hbase.client.Connection;</span> |
| <span class="source-line-no">028</span><span id="line-28">import org.apache.hadoop.hbase.client.ConnectionFactory;</span> |
| <span class="source-line-no">029</span><span id="line-29">import org.apache.hadoop.hbase.client.Delete;</span> |
| <span class="source-line-no">030</span><span id="line-30">import org.apache.hadoop.hbase.client.Durability;</span> |
| <span class="source-line-no">031</span><span id="line-31">import org.apache.hadoop.hbase.client.Mutation;</span> |
| <span class="source-line-no">032</span><span id="line-32">import org.apache.hadoop.hbase.client.Put;</span> |
| <span class="source-line-no">033</span><span id="line-33">import org.apache.hadoop.hbase.io.ImmutableBytesWritable;</span> |
| <span class="source-line-no">034</span><span id="line-34">import org.apache.hadoop.hbase.util.Bytes;</span> |
| <span class="source-line-no">035</span><span id="line-35">import org.apache.hadoop.mapreduce.JobContext;</span> |
| <span class="source-line-no">036</span><span id="line-36">import org.apache.hadoop.mapreduce.OutputCommitter;</span> |
| <span class="source-line-no">037</span><span id="line-37">import org.apache.hadoop.mapreduce.OutputFormat;</span> |
| <span class="source-line-no">038</span><span id="line-38">import org.apache.hadoop.mapreduce.RecordWriter;</span> |
| <span class="source-line-no">039</span><span id="line-39">import org.apache.hadoop.mapreduce.TaskAttemptContext;</span> |
| <span class="source-line-no">040</span><span id="line-40">import org.apache.yetus.audience.InterfaceAudience;</span> |
| <span class="source-line-no">041</span><span id="line-41">import org.slf4j.Logger;</span> |
| <span class="source-line-no">042</span><span id="line-42">import org.slf4j.LoggerFactory;</span> |
| <span class="source-line-no">043</span><span id="line-43"></span> |
| <span class="source-line-no">044</span><span id="line-44">/**</span> |
| <span class="source-line-no">045</span><span id="line-45"> * <p></span> |
| <span class="source-line-no">046</span><span id="line-46"> * Hadoop output format that writes to one or more HBase tables. The key is taken to be the table</span> |
| <span class="source-line-no">047</span><span id="line-47"> * name while the output value <em>must</em> be either a {@link Put} or a {@link Delete} instance.</span> |
| <span class="source-line-no">048</span><span id="line-48"> * All tables must already exist, and all Puts and Deletes must reference only valid column</span> |
| <span class="source-line-no">049</span><span id="line-49"> * families.</span> |
| <span class="source-line-no">050</span><span id="line-50"> * </p></span> |
| <span class="source-line-no">051</span><span id="line-51"> * <p></span> |
| <span class="source-line-no">052</span><span id="line-52"> * Write-ahead logging (WAL) for Puts can be disabled by setting {@link #WAL_PROPERTY} to</span> |
| <span class="source-line-no">053</span><span id="line-53"> * {@link #WAL_OFF}. Default value is {@link #WAL_ON}. Note that disabling write-ahead logging is</span> |
| <span class="source-line-no">054</span><span id="line-54"> * only appropriate for jobs where loss of data due to region server failure can be tolerated (for</span> |
| <span class="source-line-no">055</span><span id="line-55"> * example, because it is easy to rerun a bulk import).</span> |
| <span class="source-line-no">056</span><span id="line-56"> * </p></span> |
| <span class="source-line-no">057</span><span id="line-57"> */</span> |
| <span class="source-line-no">058</span><span id="line-58">@InterfaceAudience.Public</span> |
| <span class="source-line-no">059</span><span id="line-59">public class MultiTableOutputFormat extends OutputFormat<ImmutableBytesWritable, Mutation> {</span> |
| <span class="source-line-no">060</span><span id="line-60"> /** Set this to {@link #WAL_OFF} to turn off write-ahead logging (WAL) */</span> |
| <span class="source-line-no">061</span><span id="line-61"> public static final String WAL_PROPERTY = "hbase.mapreduce.multitableoutputformat.wal";</span> |
| <span class="source-line-no">062</span><span id="line-62"> /** Property value to use write-ahead logging */</span> |
| <span class="source-line-no">063</span><span id="line-63"> public static final boolean WAL_ON = true;</span> |
| <span class="source-line-no">064</span><span id="line-64"> /** Property value to disable write-ahead logging */</span> |
| <span class="source-line-no">065</span><span id="line-65"> public static final boolean WAL_OFF = false;</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"> * Record writer for outputting to multiple HTables.</span> |
| <span class="source-line-no">069</span><span id="line-69"> */</span> |
| <span class="source-line-no">070</span><span id="line-70"> protected static class MultiTableRecordWriter</span> |
| <span class="source-line-no">071</span><span id="line-71"> extends RecordWriter<ImmutableBytesWritable, Mutation> {</span> |
| <span class="source-line-no">072</span><span id="line-72"> private static final Logger LOG = LoggerFactory.getLogger(MultiTableRecordWriter.class);</span> |
| <span class="source-line-no">073</span><span id="line-73"> Connection connection;</span> |
| <span class="source-line-no">074</span><span id="line-74"> Map<ImmutableBytesWritable, BufferedMutator> mutatorMap = new HashMap<>();</span> |
| <span class="source-line-no">075</span><span id="line-75"> Configuration conf;</span> |
| <span class="source-line-no">076</span><span id="line-76"> boolean useWriteAheadLogging;</span> |
| <span class="source-line-no">077</span><span id="line-77"></span> |
| <span class="source-line-no">078</span><span id="line-78"> /**</span> |
| <span class="source-line-no">079</span><span id="line-79"> * HBaseConfiguration to used whether to use write ahead logging. This can be turned off (</span> |
| <span class="source-line-no">080</span><span id="line-80"> * <tt>false</tt>) to improve performance when bulk loading data.</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 MultiTableRecordWriter(Configuration conf, boolean useWriteAheadLogging)</span> |
| <span class="source-line-no">083</span><span id="line-83"> throws IOException {</span> |
| <span class="source-line-no">084</span><span id="line-84"> LOG.debug(</span> |
| <span class="source-line-no">085</span><span id="line-85"> "Created new MultiTableRecordReader with WAL " + (useWriteAheadLogging ? "on" : "off"));</span> |
| <span class="source-line-no">086</span><span id="line-86"> this.conf = conf;</span> |
| <span class="source-line-no">087</span><span id="line-87"> this.useWriteAheadLogging = useWriteAheadLogging;</span> |
| <span class="source-line-no">088</span><span id="line-88"> }</span> |
| <span class="source-line-no">089</span><span id="line-89"></span> |
| <span class="source-line-no">090</span><span id="line-90"> /**</span> |
| <span class="source-line-no">091</span><span id="line-91"> * the name of the table, as a string</span> |
| <span class="source-line-no">092</span><span id="line-92"> * @return the named mutator if there is a problem opening a table</span> |
| <span class="source-line-no">093</span><span id="line-93"> */</span> |
| <span class="source-line-no">094</span><span id="line-94"> BufferedMutator getBufferedMutator(ImmutableBytesWritable tableName) throws IOException {</span> |
| <span class="source-line-no">095</span><span id="line-95"> if (this.connection == null) {</span> |
| <span class="source-line-no">096</span><span id="line-96"> this.connection = ConnectionFactory.createConnection(conf);</span> |
| <span class="source-line-no">097</span><span id="line-97"> }</span> |
| <span class="source-line-no">098</span><span id="line-98"> if (!mutatorMap.containsKey(tableName)) {</span> |
| <span class="source-line-no">099</span><span id="line-99"> LOG.debug("Opening HTable \"" + Bytes.toString(tableName.get()) + "\" for writing");</span> |
| <span class="source-line-no">100</span><span id="line-100"></span> |
| <span class="source-line-no">101</span><span id="line-101"> BufferedMutator mutator = connection.getBufferedMutator(TableName.valueOf(tableName.get()));</span> |
| <span class="source-line-no">102</span><span id="line-102"> mutatorMap.put(tableName, mutator);</span> |
| <span class="source-line-no">103</span><span id="line-103"> }</span> |
| <span class="source-line-no">104</span><span id="line-104"> return mutatorMap.get(tableName);</span> |
| <span class="source-line-no">105</span><span id="line-105"> }</span> |
| <span class="source-line-no">106</span><span id="line-106"></span> |
| <span class="source-line-no">107</span><span id="line-107"> @Override</span> |
| <span class="source-line-no">108</span><span id="line-108"> public void close(TaskAttemptContext context) throws IOException {</span> |
| <span class="source-line-no">109</span><span id="line-109"> for (BufferedMutator mutator : mutatorMap.values()) {</span> |
| <span class="source-line-no">110</span><span id="line-110"> mutator.close();</span> |
| <span class="source-line-no">111</span><span id="line-111"> }</span> |
| <span class="source-line-no">112</span><span id="line-112"> if (connection != null) {</span> |
| <span class="source-line-no">113</span><span id="line-113"> connection.close();</span> |
| <span class="source-line-no">114</span><span id="line-114"> }</span> |
| <span class="source-line-no">115</span><span id="line-115"> }</span> |
| <span class="source-line-no">116</span><span id="line-116"></span> |
| <span class="source-line-no">117</span><span id="line-117"> /**</span> |
| <span class="source-line-no">118</span><span id="line-118"> * Writes an action (Put or Delete) to the specified table. the table being updated. the update,</span> |
| <span class="source-line-no">119</span><span id="line-119"> * either a put or a delete. if the action is not a put or a delete.</span> |
| <span class="source-line-no">120</span><span id="line-120"> */</span> |
| <span class="source-line-no">121</span><span id="line-121"> @Override</span> |
| <span class="source-line-no">122</span><span id="line-122"> public void write(ImmutableBytesWritable tableName, Mutation action) throws IOException {</span> |
| <span class="source-line-no">123</span><span id="line-123"> BufferedMutator mutator = getBufferedMutator(tableName);</span> |
| <span class="source-line-no">124</span><span id="line-124"> // The actions are not immutable, so we defensively copy them</span> |
| <span class="source-line-no">125</span><span id="line-125"> if (action instanceof Put) {</span> |
| <span class="source-line-no">126</span><span id="line-126"> Put put = new Put((Put) action);</span> |
| <span class="source-line-no">127</span><span id="line-127"> put.setDurability(useWriteAheadLogging ? Durability.SYNC_WAL : Durability.SKIP_WAL);</span> |
| <span class="source-line-no">128</span><span id="line-128"> mutator.mutate(put);</span> |
| <span class="source-line-no">129</span><span id="line-129"> } else if (action instanceof Delete) {</span> |
| <span class="source-line-no">130</span><span id="line-130"> Delete delete = new Delete((Delete) action);</span> |
| <span class="source-line-no">131</span><span id="line-131"> mutator.mutate(delete);</span> |
| <span class="source-line-no">132</span><span id="line-132"> } else throw new IllegalArgumentException("action must be either Delete or Put");</span> |
| <span class="source-line-no">133</span><span id="line-133"> }</span> |
| <span class="source-line-no">134</span><span id="line-134"> }</span> |
| <span class="source-line-no">135</span><span id="line-135"></span> |
| <span class="source-line-no">136</span><span id="line-136"> @Override</span> |
| <span class="source-line-no">137</span><span id="line-137"> public void checkOutputSpecs(JobContext context) throws IOException, InterruptedException {</span> |
| <span class="source-line-no">138</span><span id="line-138"> // we can't know ahead of time if it's going to blow up when the user</span> |
| <span class="source-line-no">139</span><span id="line-139"> // passes a table name that doesn't exist, so nothing useful here.</span> |
| <span class="source-line-no">140</span><span id="line-140"> }</span> |
| <span class="source-line-no">141</span><span id="line-141"></span> |
| <span class="source-line-no">142</span><span id="line-142"> @Override</span> |
| <span class="source-line-no">143</span><span id="line-143"> public OutputCommitter getOutputCommitter(TaskAttemptContext context)</span> |
| <span class="source-line-no">144</span><span id="line-144"> throws IOException, InterruptedException {</span> |
| <span class="source-line-no">145</span><span id="line-145"> return new TableOutputCommitter();</span> |
| <span class="source-line-no">146</span><span id="line-146"> }</span> |
| <span class="source-line-no">147</span><span id="line-147"></span> |
| <span class="source-line-no">148</span><span id="line-148"> @Override</span> |
| <span class="source-line-no">149</span><span id="line-149"> public RecordWriter<ImmutableBytesWritable, Mutation> getRecordWriter(TaskAttemptContext context)</span> |
| <span class="source-line-no">150</span><span id="line-150"> throws IOException, InterruptedException {</span> |
| <span class="source-line-no">151</span><span id="line-151"> Configuration conf = context.getConfiguration();</span> |
| <span class="source-line-no">152</span><span id="line-152"> return new MultiTableRecordWriter(HBaseConfiguration.create(conf),</span> |
| <span class="source-line-no">153</span><span id="line-153"> conf.getBoolean(WAL_PROPERTY, WAL_ON));</span> |
| <span class="source-line-no">154</span><span id="line-154"> }</span> |
| <span class="source-line-no">155</span><span id="line-155"></span> |
| <span class="source-line-no">156</span><span id="line-156">}</span> |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| </pre> |
| </div> |
| </main> |
| </body> |
| </html> |