| <!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.io.InputStreamReader;<a name="line.21"></a> |
| <span class="sourceLineNo">022</span>import java.io.OutputStreamWriter;<a name="line.22"></a> |
| <span class="sourceLineNo">023</span>import java.security.MessageDigest;<a name="line.23"></a> |
| <span class="sourceLineNo">024</span>import java.security.NoSuchAlgorithmException;<a name="line.24"></a> |
| <span class="sourceLineNo">025</span>import java.util.ArrayList;<a name="line.25"></a> |
| <span class="sourceLineNo">026</span>import java.util.Collections;<a name="line.26"></a> |
| <span class="sourceLineNo">027</span>import java.util.List;<a name="line.27"></a> |
| <span class="sourceLineNo">028</span>import java.util.Properties;<a name="line.28"></a> |
| <span class="sourceLineNo">029</span>import org.apache.hadoop.conf.Configuration;<a name="line.29"></a> |
| <span class="sourceLineNo">030</span>import org.apache.hadoop.conf.Configured;<a name="line.30"></a> |
| <span class="sourceLineNo">031</span>import org.apache.hadoop.fs.FSDataInputStream;<a name="line.31"></a> |
| <span class="sourceLineNo">032</span>import org.apache.hadoop.fs.FileSystem;<a name="line.32"></a> |
| <span class="sourceLineNo">033</span>import org.apache.hadoop.fs.Path;<a name="line.33"></a> |
| <span class="sourceLineNo">034</span>import org.apache.hadoop.hbase.Cell;<a name="line.34"></a> |
| <span class="sourceLineNo">035</span>import org.apache.hadoop.hbase.HBaseConfiguration;<a name="line.35"></a> |
| <span class="sourceLineNo">036</span>import org.apache.hadoop.hbase.HConstants;<a name="line.36"></a> |
| <span class="sourceLineNo">037</span>import org.apache.hadoop.hbase.TableName;<a name="line.37"></a> |
| <span class="sourceLineNo">038</span>import org.apache.hadoop.hbase.client.Connection;<a name="line.38"></a> |
| <span class="sourceLineNo">039</span>import org.apache.hadoop.hbase.client.ConnectionFactory;<a name="line.39"></a> |
| <span class="sourceLineNo">040</span>import org.apache.hadoop.hbase.client.Result;<a name="line.40"></a> |
| <span class="sourceLineNo">041</span>import org.apache.hadoop.hbase.client.Scan;<a name="line.41"></a> |
| <span class="sourceLineNo">042</span>import org.apache.hadoop.hbase.io.ImmutableBytesWritable;<a name="line.42"></a> |
| <span class="sourceLineNo">043</span>import org.apache.hadoop.hbase.util.Bytes;<a name="line.43"></a> |
| <span class="sourceLineNo">044</span>import org.apache.hadoop.hbase.util.Pair;<a name="line.44"></a> |
| <span class="sourceLineNo">045</span>import org.apache.hadoop.io.MapFile;<a name="line.45"></a> |
| <span class="sourceLineNo">046</span>import org.apache.hadoop.io.NullWritable;<a name="line.46"></a> |
| <span class="sourceLineNo">047</span>import org.apache.hadoop.io.SequenceFile;<a name="line.47"></a> |
| <span class="sourceLineNo">048</span>import org.apache.hadoop.mapreduce.Job;<a name="line.48"></a> |
| <span class="sourceLineNo">049</span>import org.apache.hadoop.mapreduce.Reducer;<a name="line.49"></a> |
| <span class="sourceLineNo">050</span>import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat;<a name="line.50"></a> |
| <span class="sourceLineNo">051</span>import org.apache.hadoop.mapreduce.lib.output.MapFileOutputFormat;<a name="line.51"></a> |
| <span class="sourceLineNo">052</span>import org.apache.hadoop.mapreduce.lib.partition.TotalOrderPartitioner;<a name="line.52"></a> |
| <span class="sourceLineNo">053</span>import org.apache.hadoop.util.GenericOptionsParser;<a name="line.53"></a> |
| <span class="sourceLineNo">054</span>import org.apache.hadoop.util.Tool;<a name="line.54"></a> |
| <span class="sourceLineNo">055</span>import org.apache.hadoop.util.ToolRunner;<a name="line.55"></a> |
| <span class="sourceLineNo">056</span>import org.apache.yetus.audience.InterfaceAudience;<a name="line.56"></a> |
| <span class="sourceLineNo">057</span>import org.slf4j.Logger;<a name="line.57"></a> |
| <span class="sourceLineNo">058</span>import org.slf4j.LoggerFactory;<a name="line.58"></a> |
| <span class="sourceLineNo">059</span><a name="line.59"></a> |
| <span class="sourceLineNo">060</span>import org.apache.hbase.thirdparty.com.google.common.base.Charsets;<a name="line.60"></a> |
| <span class="sourceLineNo">061</span>import org.apache.hbase.thirdparty.com.google.common.base.Throwables;<a name="line.61"></a> |
| <span class="sourceLineNo">062</span>import org.apache.hbase.thirdparty.com.google.common.collect.Ordering;<a name="line.62"></a> |
| <span class="sourceLineNo">063</span><a name="line.63"></a> |
| <span class="sourceLineNo">064</span>@InterfaceAudience.Private<a name="line.64"></a> |
| <span class="sourceLineNo">065</span>public class HashTable extends Configured implements Tool {<a name="line.65"></a> |
| <span class="sourceLineNo">066</span><a name="line.66"></a> |
| <span class="sourceLineNo">067</span> private static final Logger LOG = LoggerFactory.getLogger(HashTable.class);<a name="line.67"></a> |
| <span class="sourceLineNo">068</span><a name="line.68"></a> |
| <span class="sourceLineNo">069</span> private static final int DEFAULT_BATCH_SIZE = 8000;<a name="line.69"></a> |
| <span class="sourceLineNo">070</span><a name="line.70"></a> |
| <span class="sourceLineNo">071</span> private final static String HASH_BATCH_SIZE_CONF_KEY = "hash.batch.size";<a name="line.71"></a> |
| <span class="sourceLineNo">072</span> final static String PARTITIONS_FILE_NAME = "partitions";<a name="line.72"></a> |
| <span class="sourceLineNo">073</span> final static String MANIFEST_FILE_NAME = "manifest";<a name="line.73"></a> |
| <span class="sourceLineNo">074</span> final static String HASH_DATA_DIR = "hashes";<a name="line.74"></a> |
| <span class="sourceLineNo">075</span> final static String OUTPUT_DATA_FILE_PREFIX = "part-r-";<a name="line.75"></a> |
| <span class="sourceLineNo">076</span> final static String IGNORE_TIMESTAMPS = "ignoreTimestamps";<a name="line.76"></a> |
| <span class="sourceLineNo">077</span> private final static String TMP_MANIFEST_FILE_NAME = "manifest.tmp";<a name="line.77"></a> |
| <span class="sourceLineNo">078</span><a name="line.78"></a> |
| <span class="sourceLineNo">079</span> TableHash tableHash = new TableHash();<a name="line.79"></a> |
| <span class="sourceLineNo">080</span> Path destPath;<a name="line.80"></a> |
| <span class="sourceLineNo">081</span><a name="line.81"></a> |
| <span class="sourceLineNo">082</span> public HashTable(Configuration conf) {<a name="line.82"></a> |
| <span class="sourceLineNo">083</span> super(conf);<a name="line.83"></a> |
| <span class="sourceLineNo">084</span> }<a name="line.84"></a> |
| <span class="sourceLineNo">085</span><a name="line.85"></a> |
| <span class="sourceLineNo">086</span> public static class TableHash {<a name="line.86"></a> |
| <span class="sourceLineNo">087</span><a name="line.87"></a> |
| <span class="sourceLineNo">088</span> Path hashDir;<a name="line.88"></a> |
| <span class="sourceLineNo">089</span><a name="line.89"></a> |
| <span class="sourceLineNo">090</span> String tableName;<a name="line.90"></a> |
| <span class="sourceLineNo">091</span> String families = null;<a name="line.91"></a> |
| <span class="sourceLineNo">092</span> long batchSize = DEFAULT_BATCH_SIZE;<a name="line.92"></a> |
| <span class="sourceLineNo">093</span> int numHashFiles = 0;<a name="line.93"></a> |
| <span class="sourceLineNo">094</span> byte[] startRow = HConstants.EMPTY_START_ROW;<a name="line.94"></a> |
| <span class="sourceLineNo">095</span> byte[] stopRow = HConstants.EMPTY_END_ROW;<a name="line.95"></a> |
| <span class="sourceLineNo">096</span> int scanBatch = 0;<a name="line.96"></a> |
| <span class="sourceLineNo">097</span> int versions = -1;<a name="line.97"></a> |
| <span class="sourceLineNo">098</span> long startTime = 0;<a name="line.98"></a> |
| <span class="sourceLineNo">099</span> long endTime = 0;<a name="line.99"></a> |
| <span class="sourceLineNo">100</span> boolean ignoreTimestamps;<a name="line.100"></a> |
| <span class="sourceLineNo">101</span> boolean rawScan;<a name="line.101"></a> |
| <span class="sourceLineNo">102</span><a name="line.102"></a> |
| <span class="sourceLineNo">103</span> List<ImmutableBytesWritable> partitions;<a name="line.103"></a> |
| <span class="sourceLineNo">104</span><a name="line.104"></a> |
| <span class="sourceLineNo">105</span> public static TableHash read(Configuration conf, Path hashDir) throws IOException {<a name="line.105"></a> |
| <span class="sourceLineNo">106</span> TableHash tableHash = new TableHash();<a name="line.106"></a> |
| <span class="sourceLineNo">107</span> FileSystem fs = hashDir.getFileSystem(conf);<a name="line.107"></a> |
| <span class="sourceLineNo">108</span> tableHash.hashDir = hashDir;<a name="line.108"></a> |
| <span class="sourceLineNo">109</span> tableHash.readPropertiesFile(fs, new Path(hashDir, MANIFEST_FILE_NAME));<a name="line.109"></a> |
| <span class="sourceLineNo">110</span> tableHash.readPartitionFile(fs, conf, new Path(hashDir, PARTITIONS_FILE_NAME));<a name="line.110"></a> |
| <span class="sourceLineNo">111</span> return tableHash;<a name="line.111"></a> |
| <span class="sourceLineNo">112</span> }<a name="line.112"></a> |
| <span class="sourceLineNo">113</span><a name="line.113"></a> |
| <span class="sourceLineNo">114</span> void writePropertiesFile(FileSystem fs, Path path) throws IOException {<a name="line.114"></a> |
| <span class="sourceLineNo">115</span> Properties p = new Properties();<a name="line.115"></a> |
| <span class="sourceLineNo">116</span> p.setProperty("table", tableName);<a name="line.116"></a> |
| <span class="sourceLineNo">117</span> if (families != null) {<a name="line.117"></a> |
| <span class="sourceLineNo">118</span> p.setProperty("columnFamilies", families);<a name="line.118"></a> |
| <span class="sourceLineNo">119</span> }<a name="line.119"></a> |
| <span class="sourceLineNo">120</span> p.setProperty("targetBatchSize", Long.toString(batchSize));<a name="line.120"></a> |
| <span class="sourceLineNo">121</span> p.setProperty("numHashFiles", Integer.toString(numHashFiles));<a name="line.121"></a> |
| <span class="sourceLineNo">122</span> if (!isTableStartRow(startRow)) {<a name="line.122"></a> |
| <span class="sourceLineNo">123</span> p.setProperty("startRowHex", Bytes.toHex(startRow));<a name="line.123"></a> |
| <span class="sourceLineNo">124</span> }<a name="line.124"></a> |
| <span class="sourceLineNo">125</span> if (!isTableEndRow(stopRow)) {<a name="line.125"></a> |
| <span class="sourceLineNo">126</span> p.setProperty("stopRowHex", Bytes.toHex(stopRow));<a name="line.126"></a> |
| <span class="sourceLineNo">127</span> }<a name="line.127"></a> |
| <span class="sourceLineNo">128</span> if (scanBatch > 0) {<a name="line.128"></a> |
| <span class="sourceLineNo">129</span> p.setProperty("scanBatch", Integer.toString(scanBatch));<a name="line.129"></a> |
| <span class="sourceLineNo">130</span> }<a name="line.130"></a> |
| <span class="sourceLineNo">131</span> if (versions >= 0) {<a name="line.131"></a> |
| <span class="sourceLineNo">132</span> p.setProperty("versions", Integer.toString(versions));<a name="line.132"></a> |
| <span class="sourceLineNo">133</span> }<a name="line.133"></a> |
| <span class="sourceLineNo">134</span> if (startTime != 0) {<a name="line.134"></a> |
| <span class="sourceLineNo">135</span> p.setProperty("startTimestamp", Long.toString(startTime));<a name="line.135"></a> |
| <span class="sourceLineNo">136</span> }<a name="line.136"></a> |
| <span class="sourceLineNo">137</span> if (endTime != 0) {<a name="line.137"></a> |
| <span class="sourceLineNo">138</span> p.setProperty("endTimestamp", Long.toString(endTime));<a name="line.138"></a> |
| <span class="sourceLineNo">139</span> }<a name="line.139"></a> |
| <span class="sourceLineNo">140</span> p.setProperty("rawScan", Boolean.toString(rawScan));<a name="line.140"></a> |
| <span class="sourceLineNo">141</span><a name="line.141"></a> |
| <span class="sourceLineNo">142</span> try (OutputStreamWriter osw = new OutputStreamWriter(fs.create(path), Charsets.UTF_8)) {<a name="line.142"></a> |
| <span class="sourceLineNo">143</span> p.store(osw, null);<a name="line.143"></a> |
| <span class="sourceLineNo">144</span> }<a name="line.144"></a> |
| <span class="sourceLineNo">145</span> }<a name="line.145"></a> |
| <span class="sourceLineNo">146</span><a name="line.146"></a> |
| <span class="sourceLineNo">147</span> void readPropertiesFile(FileSystem fs, Path path) throws IOException {<a name="line.147"></a> |
| <span class="sourceLineNo">148</span> Properties p = new Properties();<a name="line.148"></a> |
| <span class="sourceLineNo">149</span> try (FSDataInputStream in = fs.open(path)) {<a name="line.149"></a> |
| <span class="sourceLineNo">150</span> try (InputStreamReader isr = new InputStreamReader(in, Charsets.UTF_8)) {<a name="line.150"></a> |
| <span class="sourceLineNo">151</span> p.load(isr);<a name="line.151"></a> |
| <span class="sourceLineNo">152</span> }<a name="line.152"></a> |
| <span class="sourceLineNo">153</span> }<a name="line.153"></a> |
| <span class="sourceLineNo">154</span> tableName = p.getProperty("table");<a name="line.154"></a> |
| <span class="sourceLineNo">155</span> families = p.getProperty("columnFamilies");<a name="line.155"></a> |
| <span class="sourceLineNo">156</span> batchSize = Long.parseLong(p.getProperty("targetBatchSize"));<a name="line.156"></a> |
| <span class="sourceLineNo">157</span> numHashFiles = Integer.parseInt(p.getProperty("numHashFiles"));<a name="line.157"></a> |
| <span class="sourceLineNo">158</span><a name="line.158"></a> |
| <span class="sourceLineNo">159</span> String startRowHex = p.getProperty("startRowHex");<a name="line.159"></a> |
| <span class="sourceLineNo">160</span> if (startRowHex != null) {<a name="line.160"></a> |
| <span class="sourceLineNo">161</span> startRow = Bytes.fromHex(startRowHex);<a name="line.161"></a> |
| <span class="sourceLineNo">162</span> }<a name="line.162"></a> |
| <span class="sourceLineNo">163</span> String stopRowHex = p.getProperty("stopRowHex");<a name="line.163"></a> |
| <span class="sourceLineNo">164</span> if (stopRowHex != null) {<a name="line.164"></a> |
| <span class="sourceLineNo">165</span> stopRow = Bytes.fromHex(stopRowHex);<a name="line.165"></a> |
| <span class="sourceLineNo">166</span> }<a name="line.166"></a> |
| <span class="sourceLineNo">167</span><a name="line.167"></a> |
| <span class="sourceLineNo">168</span> String scanBatchString = p.getProperty("scanBatch");<a name="line.168"></a> |
| <span class="sourceLineNo">169</span> if (scanBatchString != null) {<a name="line.169"></a> |
| <span class="sourceLineNo">170</span> scanBatch = Integer.parseInt(scanBatchString);<a name="line.170"></a> |
| <span class="sourceLineNo">171</span> }<a name="line.171"></a> |
| <span class="sourceLineNo">172</span><a name="line.172"></a> |
| <span class="sourceLineNo">173</span> String versionString = p.getProperty("versions");<a name="line.173"></a> |
| <span class="sourceLineNo">174</span> if (versionString != null) {<a name="line.174"></a> |
| <span class="sourceLineNo">175</span> versions = Integer.parseInt(versionString);<a name="line.175"></a> |
| <span class="sourceLineNo">176</span> }<a name="line.176"></a> |
| <span class="sourceLineNo">177</span><a name="line.177"></a> |
| <span class="sourceLineNo">178</span> String rawScanString = p.getProperty("rawScan");<a name="line.178"></a> |
| <span class="sourceLineNo">179</span> if (rawScanString != null) {<a name="line.179"></a> |
| <span class="sourceLineNo">180</span> rawScan = Boolean.parseBoolean(rawScanString);<a name="line.180"></a> |
| <span class="sourceLineNo">181</span> }<a name="line.181"></a> |
| <span class="sourceLineNo">182</span><a name="line.182"></a> |
| <span class="sourceLineNo">183</span> String startTimeString = p.getProperty("startTimestamp");<a name="line.183"></a> |
| <span class="sourceLineNo">184</span> if (startTimeString != null) {<a name="line.184"></a> |
| <span class="sourceLineNo">185</span> startTime = Long.parseLong(startTimeString);<a name="line.185"></a> |
| <span class="sourceLineNo">186</span> }<a name="line.186"></a> |
| <span class="sourceLineNo">187</span><a name="line.187"></a> |
| <span class="sourceLineNo">188</span> String endTimeString = p.getProperty("endTimestamp");<a name="line.188"></a> |
| <span class="sourceLineNo">189</span> if (endTimeString != null) {<a name="line.189"></a> |
| <span class="sourceLineNo">190</span> endTime = Long.parseLong(endTimeString);<a name="line.190"></a> |
| <span class="sourceLineNo">191</span> }<a name="line.191"></a> |
| <span class="sourceLineNo">192</span> }<a name="line.192"></a> |
| <span class="sourceLineNo">193</span><a name="line.193"></a> |
| <span class="sourceLineNo">194</span> Scan initScan() throws IOException {<a name="line.194"></a> |
| <span class="sourceLineNo">195</span> Scan scan = new Scan();<a name="line.195"></a> |
| <span class="sourceLineNo">196</span> scan.setCacheBlocks(false);<a name="line.196"></a> |
| <span class="sourceLineNo">197</span> if (startTime != 0 || endTime != 0) {<a name="line.197"></a> |
| <span class="sourceLineNo">198</span> scan.setTimeRange(startTime, endTime == 0 ? HConstants.LATEST_TIMESTAMP : endTime);<a name="line.198"></a> |
| <span class="sourceLineNo">199</span> }<a name="line.199"></a> |
| <span class="sourceLineNo">200</span> if (scanBatch > 0) {<a name="line.200"></a> |
| <span class="sourceLineNo">201</span> scan.setBatch(scanBatch);<a name="line.201"></a> |
| <span class="sourceLineNo">202</span> }<a name="line.202"></a> |
| <span class="sourceLineNo">203</span> if (versions >= 0) {<a name="line.203"></a> |
| <span class="sourceLineNo">204</span> scan.readVersions(versions);<a name="line.204"></a> |
| <span class="sourceLineNo">205</span> }<a name="line.205"></a> |
| <span class="sourceLineNo">206</span> if (!isTableStartRow(startRow)) {<a name="line.206"></a> |
| <span class="sourceLineNo">207</span> scan.withStartRow(startRow);<a name="line.207"></a> |
| <span class="sourceLineNo">208</span> }<a name="line.208"></a> |
| <span class="sourceLineNo">209</span> if (!isTableEndRow(stopRow)) {<a name="line.209"></a> |
| <span class="sourceLineNo">210</span> scan.withStopRow(stopRow);<a name="line.210"></a> |
| <span class="sourceLineNo">211</span> }<a name="line.211"></a> |
| <span class="sourceLineNo">212</span> if (families != null) {<a name="line.212"></a> |
| <span class="sourceLineNo">213</span> for (String fam : families.split(",")) {<a name="line.213"></a> |
| <span class="sourceLineNo">214</span> scan.addFamily(Bytes.toBytes(fam));<a name="line.214"></a> |
| <span class="sourceLineNo">215</span> }<a name="line.215"></a> |
| <span class="sourceLineNo">216</span> }<a name="line.216"></a> |
| <span class="sourceLineNo">217</span> scan.setRaw(rawScan);<a name="line.217"></a> |
| <span class="sourceLineNo">218</span><a name="line.218"></a> |
| <span class="sourceLineNo">219</span> return scan;<a name="line.219"></a> |
| <span class="sourceLineNo">220</span> }<a name="line.220"></a> |
| <span class="sourceLineNo">221</span><a name="line.221"></a> |
| <span class="sourceLineNo">222</span> /**<a name="line.222"></a> |
| <span class="sourceLineNo">223</span> * Choose partitions between row ranges to hash to a single output file Selects region<a name="line.223"></a> |
| <span class="sourceLineNo">224</span> * boundaries that fall within the scan range, and groups them into the desired number of<a name="line.224"></a> |
| <span class="sourceLineNo">225</span> * partitions.<a name="line.225"></a> |
| <span class="sourceLineNo">226</span> */<a name="line.226"></a> |
| <span class="sourceLineNo">227</span> void selectPartitions(Pair<byte[][], byte[][]> regionStartEndKeys) {<a name="line.227"></a> |
| <span class="sourceLineNo">228</span> List<byte[]> startKeys = new ArrayList<>();<a name="line.228"></a> |
| <span class="sourceLineNo">229</span> for (int i = 0; i < regionStartEndKeys.getFirst().length; i++) {<a name="line.229"></a> |
| <span class="sourceLineNo">230</span> byte[] regionStartKey = regionStartEndKeys.getFirst()[i];<a name="line.230"></a> |
| <span class="sourceLineNo">231</span> byte[] regionEndKey = regionStartEndKeys.getSecond()[i];<a name="line.231"></a> |
| <span class="sourceLineNo">232</span><a name="line.232"></a> |
| <span class="sourceLineNo">233</span> // if scan begins after this region, or starts before this region, then drop this region<a name="line.233"></a> |
| <span class="sourceLineNo">234</span> // in other words:<a name="line.234"></a> |
| <span class="sourceLineNo">235</span> // IF (scan begins before the end of this region<a name="line.235"></a> |
| <span class="sourceLineNo">236</span> // AND scan ends before the start of this region)<a name="line.236"></a> |
| <span class="sourceLineNo">237</span> // THEN include this region<a name="line.237"></a> |
| <span class="sourceLineNo">238</span> if (<a name="line.238"></a> |
| <span class="sourceLineNo">239</span> (isTableStartRow(startRow) || isTableEndRow(regionEndKey)<a name="line.239"></a> |
| <span class="sourceLineNo">240</span> || Bytes.compareTo(startRow, regionEndKey) < 0)<a name="line.240"></a> |
| <span class="sourceLineNo">241</span> && (isTableEndRow(stopRow) || isTableStartRow(regionStartKey)<a name="line.241"></a> |
| <span class="sourceLineNo">242</span> || Bytes.compareTo(stopRow, regionStartKey) > 0)<a name="line.242"></a> |
| <span class="sourceLineNo">243</span> ) {<a name="line.243"></a> |
| <span class="sourceLineNo">244</span> startKeys.add(regionStartKey);<a name="line.244"></a> |
| <span class="sourceLineNo">245</span> }<a name="line.245"></a> |
| <span class="sourceLineNo">246</span> }<a name="line.246"></a> |
| <span class="sourceLineNo">247</span><a name="line.247"></a> |
| <span class="sourceLineNo">248</span> int numRegions = startKeys.size();<a name="line.248"></a> |
| <span class="sourceLineNo">249</span> if (numHashFiles == 0) {<a name="line.249"></a> |
| <span class="sourceLineNo">250</span> numHashFiles = numRegions / 100;<a name="line.250"></a> |
| <span class="sourceLineNo">251</span> }<a name="line.251"></a> |
| <span class="sourceLineNo">252</span> if (numHashFiles == 0) {<a name="line.252"></a> |
| <span class="sourceLineNo">253</span> numHashFiles = 1;<a name="line.253"></a> |
| <span class="sourceLineNo">254</span> }<a name="line.254"></a> |
| <span class="sourceLineNo">255</span> if (numHashFiles > numRegions) {<a name="line.255"></a> |
| <span class="sourceLineNo">256</span> // can't partition within regions<a name="line.256"></a> |
| <span class="sourceLineNo">257</span> numHashFiles = numRegions;<a name="line.257"></a> |
| <span class="sourceLineNo">258</span> }<a name="line.258"></a> |
| <span class="sourceLineNo">259</span><a name="line.259"></a> |
| <span class="sourceLineNo">260</span> // choose a subset of start keys to group regions into ranges<a name="line.260"></a> |
| <span class="sourceLineNo">261</span> partitions = new ArrayList<>(numHashFiles - 1);<a name="line.261"></a> |
| <span class="sourceLineNo">262</span> // skip the first start key as it is not a partition between ranges.<a name="line.262"></a> |
| <span class="sourceLineNo">263</span> for (long i = 1; i < numHashFiles; i++) {<a name="line.263"></a> |
| <span class="sourceLineNo">264</span> int splitIndex = (int) (numRegions * i / numHashFiles);<a name="line.264"></a> |
| <span class="sourceLineNo">265</span> partitions.add(new ImmutableBytesWritable(startKeys.get(splitIndex)));<a name="line.265"></a> |
| <span class="sourceLineNo">266</span> }<a name="line.266"></a> |
| <span class="sourceLineNo">267</span> }<a name="line.267"></a> |
| <span class="sourceLineNo">268</span><a name="line.268"></a> |
| <span class="sourceLineNo">269</span> void writePartitionFile(Configuration conf, Path path) throws IOException {<a name="line.269"></a> |
| <span class="sourceLineNo">270</span> FileSystem fs = path.getFileSystem(conf);<a name="line.270"></a> |
| <span class="sourceLineNo">271</span> @SuppressWarnings("deprecation")<a name="line.271"></a> |
| <span class="sourceLineNo">272</span> SequenceFile.Writer writer =<a name="line.272"></a> |
| <span class="sourceLineNo">273</span> SequenceFile.createWriter(fs, conf, path, ImmutableBytesWritable.class, NullWritable.class);<a name="line.273"></a> |
| <span class="sourceLineNo">274</span><a name="line.274"></a> |
| <span class="sourceLineNo">275</span> for (int i = 0; i < partitions.size(); i++) {<a name="line.275"></a> |
| <span class="sourceLineNo">276</span> writer.append(partitions.get(i), NullWritable.get());<a name="line.276"></a> |
| <span class="sourceLineNo">277</span> }<a name="line.277"></a> |
| <span class="sourceLineNo">278</span> writer.close();<a name="line.278"></a> |
| <span class="sourceLineNo">279</span> }<a name="line.279"></a> |
| <span class="sourceLineNo">280</span><a name="line.280"></a> |
| <span class="sourceLineNo">281</span> private void readPartitionFile(FileSystem fs, Configuration conf, Path path)<a name="line.281"></a> |
| <span class="sourceLineNo">282</span> throws IOException {<a name="line.282"></a> |
| <span class="sourceLineNo">283</span> @SuppressWarnings("deprecation")<a name="line.283"></a> |
| <span class="sourceLineNo">284</span> SequenceFile.Reader reader = new SequenceFile.Reader(fs, path, conf);<a name="line.284"></a> |
| <span class="sourceLineNo">285</span> ImmutableBytesWritable key = new ImmutableBytesWritable();<a name="line.285"></a> |
| <span class="sourceLineNo">286</span> partitions = new ArrayList<>();<a name="line.286"></a> |
| <span class="sourceLineNo">287</span> while (reader.next(key)) {<a name="line.287"></a> |
| <span class="sourceLineNo">288</span> partitions.add(new ImmutableBytesWritable(key.copyBytes()));<a name="line.288"></a> |
| <span class="sourceLineNo">289</span> }<a name="line.289"></a> |
| <span class="sourceLineNo">290</span> reader.close();<a name="line.290"></a> |
| <span class="sourceLineNo">291</span><a name="line.291"></a> |
| <span class="sourceLineNo">292</span> if (!Ordering.natural().isOrdered(partitions)) {<a name="line.292"></a> |
| <span class="sourceLineNo">293</span> throw new IOException("Partitions are not ordered!");<a name="line.293"></a> |
| <span class="sourceLineNo">294</span> }<a name="line.294"></a> |
| <span class="sourceLineNo">295</span> }<a name="line.295"></a> |
| <span class="sourceLineNo">296</span><a name="line.296"></a> |
| <span class="sourceLineNo">297</span> @Override<a name="line.297"></a> |
| <span class="sourceLineNo">298</span> public String toString() {<a name="line.298"></a> |
| <span class="sourceLineNo">299</span> StringBuilder sb = new StringBuilder();<a name="line.299"></a> |
| <span class="sourceLineNo">300</span> sb.append("tableName=").append(tableName);<a name="line.300"></a> |
| <span class="sourceLineNo">301</span> if (families != null) {<a name="line.301"></a> |
| <span class="sourceLineNo">302</span> sb.append(", families=").append(families);<a name="line.302"></a> |
| <span class="sourceLineNo">303</span> }<a name="line.303"></a> |
| <span class="sourceLineNo">304</span> sb.append(", batchSize=").append(batchSize);<a name="line.304"></a> |
| <span class="sourceLineNo">305</span> sb.append(", numHashFiles=").append(numHashFiles);<a name="line.305"></a> |
| <span class="sourceLineNo">306</span> if (!isTableStartRow(startRow)) {<a name="line.306"></a> |
| <span class="sourceLineNo">307</span> sb.append(", startRowHex=").append(Bytes.toHex(startRow));<a name="line.307"></a> |
| <span class="sourceLineNo">308</span> }<a name="line.308"></a> |
| <span class="sourceLineNo">309</span> if (!isTableEndRow(stopRow)) {<a name="line.309"></a> |
| <span class="sourceLineNo">310</span> sb.append(", stopRowHex=").append(Bytes.toHex(stopRow));<a name="line.310"></a> |
| <span class="sourceLineNo">311</span> }<a name="line.311"></a> |
| <span class="sourceLineNo">312</span> if (scanBatch >= 0) {<a name="line.312"></a> |
| <span class="sourceLineNo">313</span> sb.append(", scanBatch=").append(scanBatch);<a name="line.313"></a> |
| <span class="sourceLineNo">314</span> }<a name="line.314"></a> |
| <span class="sourceLineNo">315</span> if (versions >= 0) {<a name="line.315"></a> |
| <span class="sourceLineNo">316</span> sb.append(", versions=").append(versions);<a name="line.316"></a> |
| <span class="sourceLineNo">317</span> }<a name="line.317"></a> |
| <span class="sourceLineNo">318</span> sb.append(", rawScan=").append(rawScan);<a name="line.318"></a> |
| <span class="sourceLineNo">319</span> if (startTime != 0) {<a name="line.319"></a> |
| <span class="sourceLineNo">320</span> sb.append("startTime=").append(startTime);<a name="line.320"></a> |
| <span class="sourceLineNo">321</span> }<a name="line.321"></a> |
| <span class="sourceLineNo">322</span> if (endTime != 0) {<a name="line.322"></a> |
| <span class="sourceLineNo">323</span> sb.append("endTime=").append(endTime);<a name="line.323"></a> |
| <span class="sourceLineNo">324</span> }<a name="line.324"></a> |
| <span class="sourceLineNo">325</span> return sb.toString();<a name="line.325"></a> |
| <span class="sourceLineNo">326</span> }<a name="line.326"></a> |
| <span class="sourceLineNo">327</span><a name="line.327"></a> |
| <span class="sourceLineNo">328</span> static String getDataFileName(int hashFileIndex) {<a name="line.328"></a> |
| <span class="sourceLineNo">329</span> return String.format(HashTable.OUTPUT_DATA_FILE_PREFIX + "%05d", hashFileIndex);<a name="line.329"></a> |
| <span class="sourceLineNo">330</span> }<a name="line.330"></a> |
| <span class="sourceLineNo">331</span><a name="line.331"></a> |
| <span class="sourceLineNo">332</span> /**<a name="line.332"></a> |
| <span class="sourceLineNo">333</span> * Open a TableHash.Reader starting at the first hash at or after the given key.<a name="line.333"></a> |
| <span class="sourceLineNo">334</span> */<a name="line.334"></a> |
| <span class="sourceLineNo">335</span> public Reader newReader(Configuration conf, ImmutableBytesWritable startKey)<a name="line.335"></a> |
| <span class="sourceLineNo">336</span> throws IOException {<a name="line.336"></a> |
| <span class="sourceLineNo">337</span> return new Reader(conf, startKey);<a name="line.337"></a> |
| <span class="sourceLineNo">338</span> }<a name="line.338"></a> |
| <span class="sourceLineNo">339</span><a name="line.339"></a> |
| <span class="sourceLineNo">340</span> public class Reader implements java.io.Closeable {<a name="line.340"></a> |
| <span class="sourceLineNo">341</span> private final Configuration conf;<a name="line.341"></a> |
| <span class="sourceLineNo">342</span><a name="line.342"></a> |
| <span class="sourceLineNo">343</span> private int hashFileIndex;<a name="line.343"></a> |
| <span class="sourceLineNo">344</span> private MapFile.Reader mapFileReader;<a name="line.344"></a> |
| <span class="sourceLineNo">345</span><a name="line.345"></a> |
| <span class="sourceLineNo">346</span> private boolean cachedNext;<a name="line.346"></a> |
| <span class="sourceLineNo">347</span> private ImmutableBytesWritable key;<a name="line.347"></a> |
| <span class="sourceLineNo">348</span> private ImmutableBytesWritable hash;<a name="line.348"></a> |
| <span class="sourceLineNo">349</span><a name="line.349"></a> |
| <span class="sourceLineNo">350</span> Reader(Configuration conf, ImmutableBytesWritable startKey) throws IOException {<a name="line.350"></a> |
| <span class="sourceLineNo">351</span> this.conf = conf;<a name="line.351"></a> |
| <span class="sourceLineNo">352</span> int partitionIndex = Collections.binarySearch(partitions, startKey);<a name="line.352"></a> |
| <span class="sourceLineNo">353</span> if (partitionIndex >= 0) {<a name="line.353"></a> |
| <span class="sourceLineNo">354</span> // if the key is equal to a partition, then go the file after that partition<a name="line.354"></a> |
| <span class="sourceLineNo">355</span> hashFileIndex = partitionIndex + 1;<a name="line.355"></a> |
| <span class="sourceLineNo">356</span> } else {<a name="line.356"></a> |
| <span class="sourceLineNo">357</span> // if the key is between partitions, then go to the file between those partitions<a name="line.357"></a> |
| <span class="sourceLineNo">358</span> hashFileIndex = -1 - partitionIndex;<a name="line.358"></a> |
| <span class="sourceLineNo">359</span> }<a name="line.359"></a> |
| <span class="sourceLineNo">360</span> openHashFile();<a name="line.360"></a> |
| <span class="sourceLineNo">361</span><a name="line.361"></a> |
| <span class="sourceLineNo">362</span> // MapFile's don't make it easy to seek() so that the subsequent next() returns<a name="line.362"></a> |
| <span class="sourceLineNo">363</span> // the desired key/value pair. So we cache it for the first call of next().<a name="line.363"></a> |
| <span class="sourceLineNo">364</span> hash = new ImmutableBytesWritable();<a name="line.364"></a> |
| <span class="sourceLineNo">365</span> key = (ImmutableBytesWritable) mapFileReader.getClosest(startKey, hash);<a name="line.365"></a> |
| <span class="sourceLineNo">366</span> if (key == null) {<a name="line.366"></a> |
| <span class="sourceLineNo">367</span> cachedNext = false;<a name="line.367"></a> |
| <span class="sourceLineNo">368</span> hash = null;<a name="line.368"></a> |
| <span class="sourceLineNo">369</span> } else {<a name="line.369"></a> |
| <span class="sourceLineNo">370</span> cachedNext = true;<a name="line.370"></a> |
| <span class="sourceLineNo">371</span> }<a name="line.371"></a> |
| <span class="sourceLineNo">372</span> }<a name="line.372"></a> |
| <span class="sourceLineNo">373</span><a name="line.373"></a> |
| <span class="sourceLineNo">374</span> /**<a name="line.374"></a> |
| <span class="sourceLineNo">375</span> * Read the next key/hash pair. Returns true if such a pair exists and false when at the end<a name="line.375"></a> |
| <span class="sourceLineNo">376</span> * of the data.<a name="line.376"></a> |
| <span class="sourceLineNo">377</span> */<a name="line.377"></a> |
| <span class="sourceLineNo">378</span> public boolean next() throws IOException {<a name="line.378"></a> |
| <span class="sourceLineNo">379</span> if (cachedNext) {<a name="line.379"></a> |
| <span class="sourceLineNo">380</span> cachedNext = false;<a name="line.380"></a> |
| <span class="sourceLineNo">381</span> return true;<a name="line.381"></a> |
| <span class="sourceLineNo">382</span> }<a name="line.382"></a> |
| <span class="sourceLineNo">383</span> key = new ImmutableBytesWritable();<a name="line.383"></a> |
| <span class="sourceLineNo">384</span> hash = new ImmutableBytesWritable();<a name="line.384"></a> |
| <span class="sourceLineNo">385</span> while (true) {<a name="line.385"></a> |
| <span class="sourceLineNo">386</span> boolean hasNext = mapFileReader.next(key, hash);<a name="line.386"></a> |
| <span class="sourceLineNo">387</span> if (hasNext) {<a name="line.387"></a> |
| <span class="sourceLineNo">388</span> return true;<a name="line.388"></a> |
| <span class="sourceLineNo">389</span> }<a name="line.389"></a> |
| <span class="sourceLineNo">390</span> hashFileIndex++;<a name="line.390"></a> |
| <span class="sourceLineNo">391</span> if (hashFileIndex < TableHash.this.numHashFiles) {<a name="line.391"></a> |
| <span class="sourceLineNo">392</span> mapFileReader.close();<a name="line.392"></a> |
| <span class="sourceLineNo">393</span> openHashFile();<a name="line.393"></a> |
| <span class="sourceLineNo">394</span> } else {<a name="line.394"></a> |
| <span class="sourceLineNo">395</span> key = null;<a name="line.395"></a> |
| <span class="sourceLineNo">396</span> hash = null;<a name="line.396"></a> |
| <span class="sourceLineNo">397</span> return false;<a name="line.397"></a> |
| <span class="sourceLineNo">398</span> }<a name="line.398"></a> |
| <span class="sourceLineNo">399</span> }<a name="line.399"></a> |
| <span class="sourceLineNo">400</span> }<a name="line.400"></a> |
| <span class="sourceLineNo">401</span><a name="line.401"></a> |
| <span class="sourceLineNo">402</span> /**<a name="line.402"></a> |
| <span class="sourceLineNo">403</span> * Get the current key<a name="line.403"></a> |
| <span class="sourceLineNo">404</span> * @return the current key or null if there is no current key<a name="line.404"></a> |
| <span class="sourceLineNo">405</span> */<a name="line.405"></a> |
| <span class="sourceLineNo">406</span> public ImmutableBytesWritable getCurrentKey() {<a name="line.406"></a> |
| <span class="sourceLineNo">407</span> return key;<a name="line.407"></a> |
| <span class="sourceLineNo">408</span> }<a name="line.408"></a> |
| <span class="sourceLineNo">409</span><a name="line.409"></a> |
| <span class="sourceLineNo">410</span> /**<a name="line.410"></a> |
| <span class="sourceLineNo">411</span> * Get the current hash<a name="line.411"></a> |
| <span class="sourceLineNo">412</span> * @return the current hash or null if there is no current hash<a name="line.412"></a> |
| <span class="sourceLineNo">413</span> */<a name="line.413"></a> |
| <span class="sourceLineNo">414</span> public ImmutableBytesWritable getCurrentHash() {<a name="line.414"></a> |
| <span class="sourceLineNo">415</span> return hash;<a name="line.415"></a> |
| <span class="sourceLineNo">416</span> }<a name="line.416"></a> |
| <span class="sourceLineNo">417</span><a name="line.417"></a> |
| <span class="sourceLineNo">418</span> private void openHashFile() throws IOException {<a name="line.418"></a> |
| <span class="sourceLineNo">419</span> if (mapFileReader != null) {<a name="line.419"></a> |
| <span class="sourceLineNo">420</span> mapFileReader.close();<a name="line.420"></a> |
| <span class="sourceLineNo">421</span> }<a name="line.421"></a> |
| <span class="sourceLineNo">422</span> Path dataDir = new Path(TableHash.this.hashDir, HASH_DATA_DIR);<a name="line.422"></a> |
| <span class="sourceLineNo">423</span> Path dataFile = new Path(dataDir, getDataFileName(hashFileIndex));<a name="line.423"></a> |
| <span class="sourceLineNo">424</span> mapFileReader = new MapFile.Reader(dataFile, conf);<a name="line.424"></a> |
| <span class="sourceLineNo">425</span> }<a name="line.425"></a> |
| <span class="sourceLineNo">426</span><a name="line.426"></a> |
| <span class="sourceLineNo">427</span> @Override<a name="line.427"></a> |
| <span class="sourceLineNo">428</span> public void close() throws IOException {<a name="line.428"></a> |
| <span class="sourceLineNo">429</span> mapFileReader.close();<a name="line.429"></a> |
| <span class="sourceLineNo">430</span> }<a name="line.430"></a> |
| <span class="sourceLineNo">431</span> }<a name="line.431"></a> |
| <span class="sourceLineNo">432</span> }<a name="line.432"></a> |
| <span class="sourceLineNo">433</span><a name="line.433"></a> |
| <span class="sourceLineNo">434</span> static boolean isTableStartRow(byte[] row) {<a name="line.434"></a> |
| <span class="sourceLineNo">435</span> return Bytes.equals(HConstants.EMPTY_START_ROW, row);<a name="line.435"></a> |
| <span class="sourceLineNo">436</span> }<a name="line.436"></a> |
| <span class="sourceLineNo">437</span><a name="line.437"></a> |
| <span class="sourceLineNo">438</span> static boolean isTableEndRow(byte[] row) {<a name="line.438"></a> |
| <span class="sourceLineNo">439</span> return Bytes.equals(HConstants.EMPTY_END_ROW, row);<a name="line.439"></a> |
| <span class="sourceLineNo">440</span> }<a name="line.440"></a> |
| <span class="sourceLineNo">441</span><a name="line.441"></a> |
| <span class="sourceLineNo">442</span> public Job createSubmittableJob(String[] args) throws IOException {<a name="line.442"></a> |
| <span class="sourceLineNo">443</span> Path partitionsPath = new Path(destPath, PARTITIONS_FILE_NAME);<a name="line.443"></a> |
| <span class="sourceLineNo">444</span> generatePartitions(partitionsPath);<a name="line.444"></a> |
| <span class="sourceLineNo">445</span><a name="line.445"></a> |
| <span class="sourceLineNo">446</span> Job job = Job.getInstance(getConf(),<a name="line.446"></a> |
| <span class="sourceLineNo">447</span> getConf().get("mapreduce.job.name", "hashTable_" + tableHash.tableName));<a name="line.447"></a> |
| <span class="sourceLineNo">448</span> Configuration jobConf = job.getConfiguration();<a name="line.448"></a> |
| <span class="sourceLineNo">449</span> jobConf.setLong(HASH_BATCH_SIZE_CONF_KEY, tableHash.batchSize);<a name="line.449"></a> |
| <span class="sourceLineNo">450</span> jobConf.setBoolean(IGNORE_TIMESTAMPS, tableHash.ignoreTimestamps);<a name="line.450"></a> |
| <span class="sourceLineNo">451</span> job.setJarByClass(HashTable.class);<a name="line.451"></a> |
| <span class="sourceLineNo">452</span><a name="line.452"></a> |
| <span class="sourceLineNo">453</span> TableMapReduceUtil.initTableMapperJob(tableHash.tableName, tableHash.initScan(),<a name="line.453"></a> |
| <span class="sourceLineNo">454</span> HashMapper.class, ImmutableBytesWritable.class, ImmutableBytesWritable.class, job);<a name="line.454"></a> |
| <span class="sourceLineNo">455</span><a name="line.455"></a> |
| <span class="sourceLineNo">456</span> // use a TotalOrderPartitioner and reducers to group region output into hash files<a name="line.456"></a> |
| <span class="sourceLineNo">457</span> job.setPartitionerClass(TotalOrderPartitioner.class);<a name="line.457"></a> |
| <span class="sourceLineNo">458</span> TotalOrderPartitioner.setPartitionFile(jobConf, partitionsPath);<a name="line.458"></a> |
| <span class="sourceLineNo">459</span> job.setReducerClass(Reducer.class); // identity reducer<a name="line.459"></a> |
| <span class="sourceLineNo">460</span> job.setNumReduceTasks(tableHash.numHashFiles);<a name="line.460"></a> |
| <span class="sourceLineNo">461</span> job.setOutputKeyClass(ImmutableBytesWritable.class);<a name="line.461"></a> |
| <span class="sourceLineNo">462</span> job.setOutputValueClass(ImmutableBytesWritable.class);<a name="line.462"></a> |
| <span class="sourceLineNo">463</span> job.setOutputFormatClass(MapFileOutputFormat.class);<a name="line.463"></a> |
| <span class="sourceLineNo">464</span> FileOutputFormat.setOutputPath(job, new Path(destPath, HASH_DATA_DIR));<a name="line.464"></a> |
| <span class="sourceLineNo">465</span><a name="line.465"></a> |
| <span class="sourceLineNo">466</span> return job;<a name="line.466"></a> |
| <span class="sourceLineNo">467</span> }<a name="line.467"></a> |
| <span class="sourceLineNo">468</span><a name="line.468"></a> |
| <span class="sourceLineNo">469</span> private void generatePartitions(Path partitionsPath) throws IOException {<a name="line.469"></a> |
| <span class="sourceLineNo">470</span> Connection connection = ConnectionFactory.createConnection(getConf());<a name="line.470"></a> |
| <span class="sourceLineNo">471</span> Pair<byte[][], byte[][]> regionKeys =<a name="line.471"></a> |
| <span class="sourceLineNo">472</span> connection.getRegionLocator(TableName.valueOf(tableHash.tableName)).getStartEndKeys();<a name="line.472"></a> |
| <span class="sourceLineNo">473</span> connection.close();<a name="line.473"></a> |
| <span class="sourceLineNo">474</span><a name="line.474"></a> |
| <span class="sourceLineNo">475</span> tableHash.selectPartitions(regionKeys);<a name="line.475"></a> |
| <span class="sourceLineNo">476</span> LOG.info("Writing " + tableHash.partitions.size() + " partition keys to " + partitionsPath);<a name="line.476"></a> |
| <span class="sourceLineNo">477</span><a name="line.477"></a> |
| <span class="sourceLineNo">478</span> tableHash.writePartitionFile(getConf(), partitionsPath);<a name="line.478"></a> |
| <span class="sourceLineNo">479</span> }<a name="line.479"></a> |
| <span class="sourceLineNo">480</span><a name="line.480"></a> |
| <span class="sourceLineNo">481</span> static class ResultHasher {<a name="line.481"></a> |
| <span class="sourceLineNo">482</span> private MessageDigest digest;<a name="line.482"></a> |
| <span class="sourceLineNo">483</span><a name="line.483"></a> |
| <span class="sourceLineNo">484</span> private boolean batchStarted = false;<a name="line.484"></a> |
| <span class="sourceLineNo">485</span> private ImmutableBytesWritable batchStartKey;<a name="line.485"></a> |
| <span class="sourceLineNo">486</span> private ImmutableBytesWritable batchHash;<a name="line.486"></a> |
| <span class="sourceLineNo">487</span> private long batchSize = 0;<a name="line.487"></a> |
| <span class="sourceLineNo">488</span> boolean ignoreTimestamps;<a name="line.488"></a> |
| <span class="sourceLineNo">489</span><a name="line.489"></a> |
| <span class="sourceLineNo">490</span> public ResultHasher() {<a name="line.490"></a> |
| <span class="sourceLineNo">491</span> try {<a name="line.491"></a> |
| <span class="sourceLineNo">492</span> digest = MessageDigest.getInstance("MD5");<a name="line.492"></a> |
| <span class="sourceLineNo">493</span> } catch (NoSuchAlgorithmException e) {<a name="line.493"></a> |
| <span class="sourceLineNo">494</span> Throwables.propagate(e);<a name="line.494"></a> |
| <span class="sourceLineNo">495</span> }<a name="line.495"></a> |
| <span class="sourceLineNo">496</span> }<a name="line.496"></a> |
| <span class="sourceLineNo">497</span><a name="line.497"></a> |
| <span class="sourceLineNo">498</span> public void startBatch(ImmutableBytesWritable row) {<a name="line.498"></a> |
| <span class="sourceLineNo">499</span> if (batchStarted) {<a name="line.499"></a> |
| <span class="sourceLineNo">500</span> throw new RuntimeException("Cannot start new batch without finishing existing one.");<a name="line.500"></a> |
| <span class="sourceLineNo">501</span> }<a name="line.501"></a> |
| <span class="sourceLineNo">502</span> batchStarted = true;<a name="line.502"></a> |
| <span class="sourceLineNo">503</span> batchSize = 0;<a name="line.503"></a> |
| <span class="sourceLineNo">504</span> batchStartKey = row;<a name="line.504"></a> |
| <span class="sourceLineNo">505</span> batchHash = null;<a name="line.505"></a> |
| <span class="sourceLineNo">506</span> }<a name="line.506"></a> |
| <span class="sourceLineNo">507</span><a name="line.507"></a> |
| <span class="sourceLineNo">508</span> public void hashResult(Result result) {<a name="line.508"></a> |
| <span class="sourceLineNo">509</span> if (!batchStarted) {<a name="line.509"></a> |
| <span class="sourceLineNo">510</span> throw new RuntimeException("Cannot add to batch that has not been started.");<a name="line.510"></a> |
| <span class="sourceLineNo">511</span> }<a name="line.511"></a> |
| <span class="sourceLineNo">512</span> for (Cell cell : result.rawCells()) {<a name="line.512"></a> |
| <span class="sourceLineNo">513</span> int rowLength = cell.getRowLength();<a name="line.513"></a> |
| <span class="sourceLineNo">514</span> int familyLength = cell.getFamilyLength();<a name="line.514"></a> |
| <span class="sourceLineNo">515</span> int qualifierLength = cell.getQualifierLength();<a name="line.515"></a> |
| <span class="sourceLineNo">516</span> int valueLength = cell.getValueLength();<a name="line.516"></a> |
| <span class="sourceLineNo">517</span> digest.update(cell.getRowArray(), cell.getRowOffset(), rowLength);<a name="line.517"></a> |
| <span class="sourceLineNo">518</span> digest.update(cell.getFamilyArray(), cell.getFamilyOffset(), familyLength);<a name="line.518"></a> |
| <span class="sourceLineNo">519</span> digest.update(cell.getQualifierArray(), cell.getQualifierOffset(), qualifierLength);<a name="line.519"></a> |
| <span class="sourceLineNo">520</span><a name="line.520"></a> |
| <span class="sourceLineNo">521</span> if (!ignoreTimestamps) {<a name="line.521"></a> |
| <span class="sourceLineNo">522</span> long ts = cell.getTimestamp();<a name="line.522"></a> |
| <span class="sourceLineNo">523</span> for (int i = 8; i > 0; i--) {<a name="line.523"></a> |
| <span class="sourceLineNo">524</span> digest.update((byte) ts);<a name="line.524"></a> |
| <span class="sourceLineNo">525</span> ts >>>= 8;<a name="line.525"></a> |
| <span class="sourceLineNo">526</span> }<a name="line.526"></a> |
| <span class="sourceLineNo">527</span> }<a name="line.527"></a> |
| <span class="sourceLineNo">528</span> digest.update(cell.getValueArray(), cell.getValueOffset(), valueLength);<a name="line.528"></a> |
| <span class="sourceLineNo">529</span><a name="line.529"></a> |
| <span class="sourceLineNo">530</span> batchSize += rowLength + familyLength + qualifierLength + 8 + valueLength;<a name="line.530"></a> |
| <span class="sourceLineNo">531</span> }<a name="line.531"></a> |
| <span class="sourceLineNo">532</span> }<a name="line.532"></a> |
| <span class="sourceLineNo">533</span><a name="line.533"></a> |
| <span class="sourceLineNo">534</span> public void finishBatch() {<a name="line.534"></a> |
| <span class="sourceLineNo">535</span> if (!batchStarted) {<a name="line.535"></a> |
| <span class="sourceLineNo">536</span> throw new RuntimeException("Cannot finish batch that has not started.");<a name="line.536"></a> |
| <span class="sourceLineNo">537</span> }<a name="line.537"></a> |
| <span class="sourceLineNo">538</span> batchStarted = false;<a name="line.538"></a> |
| <span class="sourceLineNo">539</span> batchHash = new ImmutableBytesWritable(digest.digest());<a name="line.539"></a> |
| <span class="sourceLineNo">540</span> }<a name="line.540"></a> |
| <span class="sourceLineNo">541</span><a name="line.541"></a> |
| <span class="sourceLineNo">542</span> public boolean isBatchStarted() {<a name="line.542"></a> |
| <span class="sourceLineNo">543</span> return batchStarted;<a name="line.543"></a> |
| <span class="sourceLineNo">544</span> }<a name="line.544"></a> |
| <span class="sourceLineNo">545</span><a name="line.545"></a> |
| <span class="sourceLineNo">546</span> public ImmutableBytesWritable getBatchStartKey() {<a name="line.546"></a> |
| <span class="sourceLineNo">547</span> return batchStartKey;<a name="line.547"></a> |
| <span class="sourceLineNo">548</span> }<a name="line.548"></a> |
| <span class="sourceLineNo">549</span><a name="line.549"></a> |
| <span class="sourceLineNo">550</span> public ImmutableBytesWritable getBatchHash() {<a name="line.550"></a> |
| <span class="sourceLineNo">551</span> return batchHash;<a name="line.551"></a> |
| <span class="sourceLineNo">552</span> }<a name="line.552"></a> |
| <span class="sourceLineNo">553</span><a name="line.553"></a> |
| <span class="sourceLineNo">554</span> public long getBatchSize() {<a name="line.554"></a> |
| <span class="sourceLineNo">555</span> return batchSize;<a name="line.555"></a> |
| <span class="sourceLineNo">556</span> }<a name="line.556"></a> |
| <span class="sourceLineNo">557</span> }<a name="line.557"></a> |
| <span class="sourceLineNo">558</span><a name="line.558"></a> |
| <span class="sourceLineNo">559</span> public static class HashMapper<a name="line.559"></a> |
| <span class="sourceLineNo">560</span> extends TableMapper<ImmutableBytesWritable, ImmutableBytesWritable> {<a name="line.560"></a> |
| <span class="sourceLineNo">561</span><a name="line.561"></a> |
| <span class="sourceLineNo">562</span> private ResultHasher hasher;<a name="line.562"></a> |
| <span class="sourceLineNo">563</span> private long targetBatchSize;<a name="line.563"></a> |
| <span class="sourceLineNo">564</span><a name="line.564"></a> |
| <span class="sourceLineNo">565</span> private ImmutableBytesWritable currentRow;<a name="line.565"></a> |
| <span class="sourceLineNo">566</span><a name="line.566"></a> |
| <span class="sourceLineNo">567</span> @Override<a name="line.567"></a> |
| <span class="sourceLineNo">568</span> protected void setup(Context context) throws IOException, InterruptedException {<a name="line.568"></a> |
| <span class="sourceLineNo">569</span> targetBatchSize =<a name="line.569"></a> |
| <span class="sourceLineNo">570</span> context.getConfiguration().getLong(HASH_BATCH_SIZE_CONF_KEY, DEFAULT_BATCH_SIZE);<a name="line.570"></a> |
| <span class="sourceLineNo">571</span> hasher = new ResultHasher();<a name="line.571"></a> |
| <span class="sourceLineNo">572</span> hasher.ignoreTimestamps = context.getConfiguration().getBoolean(IGNORE_TIMESTAMPS, false);<a name="line.572"></a> |
| <span class="sourceLineNo">573</span> TableSplit split = (TableSplit) context.getInputSplit();<a name="line.573"></a> |
| <span class="sourceLineNo">574</span> hasher.startBatch(new ImmutableBytesWritable(split.getStartRow()));<a name="line.574"></a> |
| <span class="sourceLineNo">575</span> }<a name="line.575"></a> |
| <span class="sourceLineNo">576</span><a name="line.576"></a> |
| <span class="sourceLineNo">577</span> @Override<a name="line.577"></a> |
| <span class="sourceLineNo">578</span> protected void map(ImmutableBytesWritable key, Result value, Context context)<a name="line.578"></a> |
| <span class="sourceLineNo">579</span> throws IOException, InterruptedException {<a name="line.579"></a> |
| <span class="sourceLineNo">580</span><a name="line.580"></a> |
| <span class="sourceLineNo">581</span> if (currentRow == null || !currentRow.equals(key)) {<a name="line.581"></a> |
| <span class="sourceLineNo">582</span> currentRow = new ImmutableBytesWritable(key); // not immutable<a name="line.582"></a> |
| <span class="sourceLineNo">583</span><a name="line.583"></a> |
| <span class="sourceLineNo">584</span> if (hasher.getBatchSize() >= targetBatchSize) {<a name="line.584"></a> |
| <span class="sourceLineNo">585</span> hasher.finishBatch();<a name="line.585"></a> |
| <span class="sourceLineNo">586</span> context.write(hasher.getBatchStartKey(), hasher.getBatchHash());<a name="line.586"></a> |
| <span class="sourceLineNo">587</span> hasher.startBatch(currentRow);<a name="line.587"></a> |
| <span class="sourceLineNo">588</span> }<a name="line.588"></a> |
| <span class="sourceLineNo">589</span> }<a name="line.589"></a> |
| <span class="sourceLineNo">590</span><a name="line.590"></a> |
| <span class="sourceLineNo">591</span> hasher.hashResult(value);<a name="line.591"></a> |
| <span class="sourceLineNo">592</span> }<a name="line.592"></a> |
| <span class="sourceLineNo">593</span><a name="line.593"></a> |
| <span class="sourceLineNo">594</span> @Override<a name="line.594"></a> |
| <span class="sourceLineNo">595</span> protected void cleanup(Context context) throws IOException, InterruptedException {<a name="line.595"></a> |
| <span class="sourceLineNo">596</span> hasher.finishBatch();<a name="line.596"></a> |
| <span class="sourceLineNo">597</span> context.write(hasher.getBatchStartKey(), hasher.getBatchHash());<a name="line.597"></a> |
| <span class="sourceLineNo">598</span> }<a name="line.598"></a> |
| <span class="sourceLineNo">599</span> }<a name="line.599"></a> |
| <span class="sourceLineNo">600</span><a name="line.600"></a> |
| <span class="sourceLineNo">601</span> private void writeTempManifestFile() throws IOException {<a name="line.601"></a> |
| <span class="sourceLineNo">602</span> Path tempManifestPath = new Path(destPath, TMP_MANIFEST_FILE_NAME);<a name="line.602"></a> |
| <span class="sourceLineNo">603</span> FileSystem fs = tempManifestPath.getFileSystem(getConf());<a name="line.603"></a> |
| <span class="sourceLineNo">604</span> tableHash.writePropertiesFile(fs, tempManifestPath);<a name="line.604"></a> |
| <span class="sourceLineNo">605</span> }<a name="line.605"></a> |
| <span class="sourceLineNo">606</span><a name="line.606"></a> |
| <span class="sourceLineNo">607</span> private void completeManifest() throws IOException {<a name="line.607"></a> |
| <span class="sourceLineNo">608</span> Path tempManifestPath = new Path(destPath, TMP_MANIFEST_FILE_NAME);<a name="line.608"></a> |
| <span class="sourceLineNo">609</span> Path manifestPath = new Path(destPath, MANIFEST_FILE_NAME);<a name="line.609"></a> |
| <span class="sourceLineNo">610</span> FileSystem fs = tempManifestPath.getFileSystem(getConf());<a name="line.610"></a> |
| <span class="sourceLineNo">611</span> fs.rename(tempManifestPath, manifestPath);<a name="line.611"></a> |
| <span class="sourceLineNo">612</span> }<a name="line.612"></a> |
| <span class="sourceLineNo">613</span><a name="line.613"></a> |
| <span class="sourceLineNo">614</span> private static final int NUM_ARGS = 2;<a name="line.614"></a> |
| <span class="sourceLineNo">615</span><a name="line.615"></a> |
| <span class="sourceLineNo">616</span> private static void printUsage(final String errorMsg) {<a name="line.616"></a> |
| <span class="sourceLineNo">617</span> if (errorMsg != null && errorMsg.length() > 0) {<a name="line.617"></a> |
| <span class="sourceLineNo">618</span> System.err.println("ERROR: " + errorMsg);<a name="line.618"></a> |
| <span class="sourceLineNo">619</span> System.err.println();<a name="line.619"></a> |
| <span class="sourceLineNo">620</span> }<a name="line.620"></a> |
| <span class="sourceLineNo">621</span> System.err.println("Usage: HashTable [options] <tablename> <outputpath>");<a name="line.621"></a> |
| <span class="sourceLineNo">622</span> System.err.println();<a name="line.622"></a> |
| <span class="sourceLineNo">623</span> System.err.println("Options:");<a name="line.623"></a> |
| <span class="sourceLineNo">624</span> System.err.println(" batchsize the target amount of bytes to hash in each batch");<a name="line.624"></a> |
| <span class="sourceLineNo">625</span> System.err.println(" rows are added to the batch until this size is reached");<a name="line.625"></a> |
| <span class="sourceLineNo">626</span> System.err.println(" (defaults to " + DEFAULT_BATCH_SIZE + " bytes)");<a name="line.626"></a> |
| <span class="sourceLineNo">627</span> System.err.println(" numhashfiles the number of hash files to create");<a name="line.627"></a> |
| <span class="sourceLineNo">628</span> System.err.println(" if set to fewer than number of regions then");<a name="line.628"></a> |
| <span class="sourceLineNo">629</span> System.err.println(" the job will create this number of reducers");<a name="line.629"></a> |
| <span class="sourceLineNo">630</span> System.err.println(" (defaults to 1/100 of regions -- at least 1)");<a name="line.630"></a> |
| <span class="sourceLineNo">631</span> System.err.println(" startrow the start row");<a name="line.631"></a> |
| <span class="sourceLineNo">632</span> System.err.println(" stoprow the stop row");<a name="line.632"></a> |
| <span class="sourceLineNo">633</span> System.err.println(" starttime beginning of the time range (unixtime in millis)");<a name="line.633"></a> |
| <span class="sourceLineNo">634</span> System.err.println(" without endtime means from starttime to forever");<a name="line.634"></a> |
| <span class="sourceLineNo">635</span> System.err.println(" endtime end of the time range.");<a name="line.635"></a> |
| <span class="sourceLineNo">636</span> System.err.println(" Ignored if no starttime specified.");<a name="line.636"></a> |
| <span class="sourceLineNo">637</span> System.err.println(" scanbatch scanner batch size to support intra row scans");<a name="line.637"></a> |
| <span class="sourceLineNo">638</span> System.err.println(" versions number of cell versions to include");<a name="line.638"></a> |
| <span class="sourceLineNo">639</span> System.err.println(" rawScan performs a raw scan (false if omitted)");<a name="line.639"></a> |
| <span class="sourceLineNo">640</span> System.err.println(" families comma-separated list of families to include");<a name="line.640"></a> |
| <span class="sourceLineNo">641</span> System.err.println(" ignoreTimestamps if true, ignores cell timestamps");<a name="line.641"></a> |
| <span class="sourceLineNo">642</span> System.err.println(" when calculating hashes");<a name="line.642"></a> |
| <span class="sourceLineNo">643</span> System.err.println();<a name="line.643"></a> |
| <span class="sourceLineNo">644</span> System.err.println("Args:");<a name="line.644"></a> |
| <span class="sourceLineNo">645</span> System.err.println(" tablename Name of the table to hash");<a name="line.645"></a> |
| <span class="sourceLineNo">646</span> System.err.println(" outputpath Filesystem path to put the output data");<a name="line.646"></a> |
| <span class="sourceLineNo">647</span> System.err.println();<a name="line.647"></a> |
| <span class="sourceLineNo">648</span> System.err.println("Examples:");<a name="line.648"></a> |
| <span class="sourceLineNo">649</span> System.err.println(" To hash 'TestTable' in 32kB batches for a 1 hour window into 50 files:");<a name="line.649"></a> |
| <span class="sourceLineNo">650</span> System.err.println(" $ hbase "<a name="line.650"></a> |
| <span class="sourceLineNo">651</span> + "org.apache.hadoop.hbase.mapreduce.HashTable --batchsize=32000 --numhashfiles=50"<a name="line.651"></a> |
| <span class="sourceLineNo">652</span> + " --starttime=1265875194289 --endtime=1265878794289 --families=cf2,cf3"<a name="line.652"></a> |
| <span class="sourceLineNo">653</span> + " TestTable /hashes/testTable");<a name="line.653"></a> |
| <span class="sourceLineNo">654</span> }<a name="line.654"></a> |
| <span class="sourceLineNo">655</span><a name="line.655"></a> |
| <span class="sourceLineNo">656</span> private boolean doCommandLine(final String[] args) {<a name="line.656"></a> |
| <span class="sourceLineNo">657</span> if (args.length < NUM_ARGS) {<a name="line.657"></a> |
| <span class="sourceLineNo">658</span> printUsage(null);<a name="line.658"></a> |
| <span class="sourceLineNo">659</span> return false;<a name="line.659"></a> |
| <span class="sourceLineNo">660</span> }<a name="line.660"></a> |
| <span class="sourceLineNo">661</span> try {<a name="line.661"></a> |
| <span class="sourceLineNo">662</span><a name="line.662"></a> |
| <span class="sourceLineNo">663</span> tableHash.tableName = args[args.length - 2];<a name="line.663"></a> |
| <span class="sourceLineNo">664</span> destPath = new Path(args[args.length - 1]);<a name="line.664"></a> |
| <span class="sourceLineNo">665</span><a name="line.665"></a> |
| <span class="sourceLineNo">666</span> for (int i = 0; i < args.length - NUM_ARGS; i++) {<a name="line.666"></a> |
| <span class="sourceLineNo">667</span> String cmd = args[i];<a name="line.667"></a> |
| <span class="sourceLineNo">668</span> if (cmd.equals("-h") || cmd.startsWith("--h")) {<a name="line.668"></a> |
| <span class="sourceLineNo">669</span> printUsage(null);<a name="line.669"></a> |
| <span class="sourceLineNo">670</span> return false;<a name="line.670"></a> |
| <span class="sourceLineNo">671</span> }<a name="line.671"></a> |
| <span class="sourceLineNo">672</span><a name="line.672"></a> |
| <span class="sourceLineNo">673</span> final String batchSizeArgKey = "--batchsize=";<a name="line.673"></a> |
| <span class="sourceLineNo">674</span> if (cmd.startsWith(batchSizeArgKey)) {<a name="line.674"></a> |
| <span class="sourceLineNo">675</span> tableHash.batchSize = Long.parseLong(cmd.substring(batchSizeArgKey.length()));<a name="line.675"></a> |
| <span class="sourceLineNo">676</span> continue;<a name="line.676"></a> |
| <span class="sourceLineNo">677</span> }<a name="line.677"></a> |
| <span class="sourceLineNo">678</span><a name="line.678"></a> |
| <span class="sourceLineNo">679</span> final String numHashFilesArgKey = "--numhashfiles=";<a name="line.679"></a> |
| <span class="sourceLineNo">680</span> if (cmd.startsWith(numHashFilesArgKey)) {<a name="line.680"></a> |
| <span class="sourceLineNo">681</span> tableHash.numHashFiles = Integer.parseInt(cmd.substring(numHashFilesArgKey.length()));<a name="line.681"></a> |
| <span class="sourceLineNo">682</span> continue;<a name="line.682"></a> |
| <span class="sourceLineNo">683</span> }<a name="line.683"></a> |
| <span class="sourceLineNo">684</span><a name="line.684"></a> |
| <span class="sourceLineNo">685</span> final String startRowArgKey = "--startrow=";<a name="line.685"></a> |
| <span class="sourceLineNo">686</span> if (cmd.startsWith(startRowArgKey)) {<a name="line.686"></a> |
| <span class="sourceLineNo">687</span> tableHash.startRow = Bytes.fromHex(cmd.substring(startRowArgKey.length()));<a name="line.687"></a> |
| <span class="sourceLineNo">688</span> continue;<a name="line.688"></a> |
| <span class="sourceLineNo">689</span> }<a name="line.689"></a> |
| <span class="sourceLineNo">690</span><a name="line.690"></a> |
| <span class="sourceLineNo">691</span> final String stopRowArgKey = "--stoprow=";<a name="line.691"></a> |
| <span class="sourceLineNo">692</span> if (cmd.startsWith(stopRowArgKey)) {<a name="line.692"></a> |
| <span class="sourceLineNo">693</span> tableHash.stopRow = Bytes.fromHex(cmd.substring(stopRowArgKey.length()));<a name="line.693"></a> |
| <span class="sourceLineNo">694</span> continue;<a name="line.694"></a> |
| <span class="sourceLineNo">695</span> }<a name="line.695"></a> |
| <span class="sourceLineNo">696</span><a name="line.696"></a> |
| <span class="sourceLineNo">697</span> final String startTimeArgKey = "--starttime=";<a name="line.697"></a> |
| <span class="sourceLineNo">698</span> if (cmd.startsWith(startTimeArgKey)) {<a name="line.698"></a> |
| <span class="sourceLineNo">699</span> tableHash.startTime = Long.parseLong(cmd.substring(startTimeArgKey.length()));<a name="line.699"></a> |
| <span class="sourceLineNo">700</span> continue;<a name="line.700"></a> |
| <span class="sourceLineNo">701</span> }<a name="line.701"></a> |
| <span class="sourceLineNo">702</span><a name="line.702"></a> |
| <span class="sourceLineNo">703</span> final String endTimeArgKey = "--endtime=";<a name="line.703"></a> |
| <span class="sourceLineNo">704</span> if (cmd.startsWith(endTimeArgKey)) {<a name="line.704"></a> |
| <span class="sourceLineNo">705</span> tableHash.endTime = Long.parseLong(cmd.substring(endTimeArgKey.length()));<a name="line.705"></a> |
| <span class="sourceLineNo">706</span> continue;<a name="line.706"></a> |
| <span class="sourceLineNo">707</span> }<a name="line.707"></a> |
| <span class="sourceLineNo">708</span><a name="line.708"></a> |
| <span class="sourceLineNo">709</span> final String scanBatchArgKey = "--scanbatch=";<a name="line.709"></a> |
| <span class="sourceLineNo">710</span> if (cmd.startsWith(scanBatchArgKey)) {<a name="line.710"></a> |
| <span class="sourceLineNo">711</span> tableHash.scanBatch = Integer.parseInt(cmd.substring(scanBatchArgKey.length()));<a name="line.711"></a> |
| <span class="sourceLineNo">712</span> continue;<a name="line.712"></a> |
| <span class="sourceLineNo">713</span> }<a name="line.713"></a> |
| <span class="sourceLineNo">714</span><a name="line.714"></a> |
| <span class="sourceLineNo">715</span> final String versionsArgKey = "--versions=";<a name="line.715"></a> |
| <span class="sourceLineNo">716</span> if (cmd.startsWith(versionsArgKey)) {<a name="line.716"></a> |
| <span class="sourceLineNo">717</span> tableHash.versions = Integer.parseInt(cmd.substring(versionsArgKey.length()));<a name="line.717"></a> |
| <span class="sourceLineNo">718</span> continue;<a name="line.718"></a> |
| <span class="sourceLineNo">719</span> }<a name="line.719"></a> |
| <span class="sourceLineNo">720</span><a name="line.720"></a> |
| <span class="sourceLineNo">721</span> final String rawScanArgKey = "--rawScan=";<a name="line.721"></a> |
| <span class="sourceLineNo">722</span> if (cmd.startsWith(rawScanArgKey)) {<a name="line.722"></a> |
| <span class="sourceLineNo">723</span> tableHash.rawScan = Boolean.parseBoolean(cmd.substring(rawScanArgKey.length()));<a name="line.723"></a> |
| <span class="sourceLineNo">724</span> continue;<a name="line.724"></a> |
| <span class="sourceLineNo">725</span> }<a name="line.725"></a> |
| <span class="sourceLineNo">726</span><a name="line.726"></a> |
| <span class="sourceLineNo">727</span> final String familiesArgKey = "--families=";<a name="line.727"></a> |
| <span class="sourceLineNo">728</span> if (cmd.startsWith(familiesArgKey)) {<a name="line.728"></a> |
| <span class="sourceLineNo">729</span> tableHash.families = cmd.substring(familiesArgKey.length());<a name="line.729"></a> |
| <span class="sourceLineNo">730</span> continue;<a name="line.730"></a> |
| <span class="sourceLineNo">731</span> }<a name="line.731"></a> |
| <span class="sourceLineNo">732</span><a name="line.732"></a> |
| <span class="sourceLineNo">733</span> final String ignoreTimestampsKey = "--ignoreTimestamps=";<a name="line.733"></a> |
| <span class="sourceLineNo">734</span> if (cmd.startsWith(ignoreTimestampsKey)) {<a name="line.734"></a> |
| <span class="sourceLineNo">735</span> tableHash.ignoreTimestamps =<a name="line.735"></a> |
| <span class="sourceLineNo">736</span> Boolean.parseBoolean(cmd.substring(ignoreTimestampsKey.length()));<a name="line.736"></a> |
| <span class="sourceLineNo">737</span> continue;<a name="line.737"></a> |
| <span class="sourceLineNo">738</span> }<a name="line.738"></a> |
| <span class="sourceLineNo">739</span><a name="line.739"></a> |
| <span class="sourceLineNo">740</span> printUsage("Invalid argument '" + cmd + "'");<a name="line.740"></a> |
| <span class="sourceLineNo">741</span> return false;<a name="line.741"></a> |
| <span class="sourceLineNo">742</span> }<a name="line.742"></a> |
| <span class="sourceLineNo">743</span> if (<a name="line.743"></a> |
| <span class="sourceLineNo">744</span> (tableHash.startTime != 0 || tableHash.endTime != 0)<a name="line.744"></a> |
| <span class="sourceLineNo">745</span> && (tableHash.startTime >= tableHash.endTime)<a name="line.745"></a> |
| <span class="sourceLineNo">746</span> ) {<a name="line.746"></a> |
| <span class="sourceLineNo">747</span> printUsage("Invalid time range filter: starttime=" + tableHash.startTime + " >= endtime="<a name="line.747"></a> |
| <span class="sourceLineNo">748</span> + tableHash.endTime);<a name="line.748"></a> |
| <span class="sourceLineNo">749</span> return false;<a name="line.749"></a> |
| <span class="sourceLineNo">750</span> }<a name="line.750"></a> |
| <span class="sourceLineNo">751</span><a name="line.751"></a> |
| <span class="sourceLineNo">752</span> } catch (Exception e) {<a name="line.752"></a> |
| <span class="sourceLineNo">753</span> LOG.error("Failed to parse commandLine arguments", e);<a name="line.753"></a> |
| <span class="sourceLineNo">754</span> printUsage("Can't start because " + e.getMessage());<a name="line.754"></a> |
| <span class="sourceLineNo">755</span> return false;<a name="line.755"></a> |
| <span class="sourceLineNo">756</span> }<a name="line.756"></a> |
| <span class="sourceLineNo">757</span> return true;<a name="line.757"></a> |
| <span class="sourceLineNo">758</span> }<a name="line.758"></a> |
| <span class="sourceLineNo">759</span><a name="line.759"></a> |
| <span class="sourceLineNo">760</span> /**<a name="line.760"></a> |
| <span class="sourceLineNo">761</span> * Main entry point.<a name="line.761"></a> |
| <span class="sourceLineNo">762</span> */<a name="line.762"></a> |
| <span class="sourceLineNo">763</span> public static void main(String[] args) throws Exception {<a name="line.763"></a> |
| <span class="sourceLineNo">764</span> int ret = ToolRunner.run(new HashTable(HBaseConfiguration.create()), args);<a name="line.764"></a> |
| <span class="sourceLineNo">765</span> System.exit(ret);<a name="line.765"></a> |
| <span class="sourceLineNo">766</span> }<a name="line.766"></a> |
| <span class="sourceLineNo">767</span><a name="line.767"></a> |
| <span class="sourceLineNo">768</span> @Override<a name="line.768"></a> |
| <span class="sourceLineNo">769</span> public int run(String[] args) throws Exception {<a name="line.769"></a> |
| <span class="sourceLineNo">770</span> String[] otherArgs = new GenericOptionsParser(getConf(), args).getRemainingArgs();<a name="line.770"></a> |
| <span class="sourceLineNo">771</span> if (!doCommandLine(otherArgs)) {<a name="line.771"></a> |
| <span class="sourceLineNo">772</span> return 1;<a name="line.772"></a> |
| <span class="sourceLineNo">773</span> }<a name="line.773"></a> |
| <span class="sourceLineNo">774</span><a name="line.774"></a> |
| <span class="sourceLineNo">775</span> Job job = createSubmittableJob(otherArgs);<a name="line.775"></a> |
| <span class="sourceLineNo">776</span> writeTempManifestFile();<a name="line.776"></a> |
| <span class="sourceLineNo">777</span> if (!job.waitForCompletion(true)) {<a name="line.777"></a> |
| <span class="sourceLineNo">778</span> LOG.info("Map-reduce job failed!");<a name="line.778"></a> |
| <span class="sourceLineNo">779</span> return 1;<a name="line.779"></a> |
| <span class="sourceLineNo">780</span> }<a name="line.780"></a> |
| <span class="sourceLineNo">781</span> completeManifest();<a name="line.781"></a> |
| <span class="sourceLineNo">782</span> return 0;<a name="line.782"></a> |
| <span class="sourceLineNo">783</span> }<a name="line.783"></a> |
| <span class="sourceLineNo">784</span><a name="line.784"></a> |
| <span class="sourceLineNo">785</span>}<a name="line.785"></a> |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| </pre> |
| </div> |
| </body> |
| </html> |