blob: 48142e3e4b55d7f0cf202925a8f0c67a0f735a31 [file] [log] [blame]
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd">
<html lang="en">
<head>
<title>Source code</title>
<link rel="stylesheet" type="text/css" href="../../../../../../stylesheet.css" title="Style">
</head>
<body>
<div class="sourceContainer">
<pre><span class="sourceLineNo">001</span>/*<a name="line.1"></a>
<span class="sourceLineNo">002</span> * Licensed to the Apache Software Foundation (ASF) under one<a name="line.2"></a>
<span class="sourceLineNo">003</span> * or more contributor license agreements. See the NOTICE file<a name="line.3"></a>
<span class="sourceLineNo">004</span> * distributed with this work for additional information<a name="line.4"></a>
<span class="sourceLineNo">005</span> * regarding copyright ownership. The ASF licenses this file<a name="line.5"></a>
<span class="sourceLineNo">006</span> * to you under the Apache License, Version 2.0 (the<a name="line.6"></a>
<span class="sourceLineNo">007</span> * "License"); you may not use this file except in compliance<a name="line.7"></a>
<span class="sourceLineNo">008</span> * with the License. You may obtain a copy of the License at<a name="line.8"></a>
<span class="sourceLineNo">009</span> *<a name="line.9"></a>
<span class="sourceLineNo">010</span> * http://www.apache.org/licenses/LICENSE-2.0<a name="line.10"></a>
<span class="sourceLineNo">011</span> *<a name="line.11"></a>
<span class="sourceLineNo">012</span> * Unless required by applicable law or agreed to in writing, software<a name="line.12"></a>
<span class="sourceLineNo">013</span> * distributed under the License is distributed on an "AS IS" BASIS,<a name="line.13"></a>
<span class="sourceLineNo">014</span> * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.<a name="line.14"></a>
<span class="sourceLineNo">015</span> * See the License for the specific language governing permissions and<a name="line.15"></a>
<span class="sourceLineNo">016</span> * limitations under the License.<a name="line.16"></a>
<span class="sourceLineNo">017</span> */<a name="line.17"></a>
<span class="sourceLineNo">018</span>package org.apache.hadoop.hbase.mapreduce;<a name="line.18"></a>
<span class="sourceLineNo">019</span><a name="line.19"></a>
<span class="sourceLineNo">020</span>import static org.junit.Assert.assertEquals;<a name="line.20"></a>
<span class="sourceLineNo">021</span>import static org.junit.Assert.assertFalse;<a name="line.21"></a>
<span class="sourceLineNo">022</span>import static org.junit.Assert.assertTrue;<a name="line.22"></a>
<span class="sourceLineNo">023</span><a name="line.23"></a>
<span class="sourceLineNo">024</span>import java.io.IOException;<a name="line.24"></a>
<span class="sourceLineNo">025</span>import java.util.Arrays;<a name="line.25"></a>
<span class="sourceLineNo">026</span>import java.util.HashMap;<a name="line.26"></a>
<span class="sourceLineNo">027</span>import java.util.HashSet;<a name="line.27"></a>
<span class="sourceLineNo">028</span>import java.util.Iterator;<a name="line.28"></a>
<span class="sourceLineNo">029</span>import java.util.List;<a name="line.29"></a>
<span class="sourceLineNo">030</span>import java.util.Map;<a name="line.30"></a>
<span class="sourceLineNo">031</span>import java.util.Set;<a name="line.31"></a>
<span class="sourceLineNo">032</span>import org.apache.hadoop.conf.Configurable;<a name="line.32"></a>
<span class="sourceLineNo">033</span>import org.apache.hadoop.conf.Configuration;<a name="line.33"></a>
<span class="sourceLineNo">034</span>import org.apache.hadoop.fs.FSDataOutputStream;<a name="line.34"></a>
<span class="sourceLineNo">035</span>import org.apache.hadoop.fs.FileStatus;<a name="line.35"></a>
<span class="sourceLineNo">036</span>import org.apache.hadoop.fs.FileSystem;<a name="line.36"></a>
<span class="sourceLineNo">037</span>import org.apache.hadoop.fs.Path;<a name="line.37"></a>
<span class="sourceLineNo">038</span>import org.apache.hadoop.hbase.Cell;<a name="line.38"></a>
<span class="sourceLineNo">039</span>import org.apache.hadoop.hbase.CellUtil;<a name="line.39"></a>
<span class="sourceLineNo">040</span>import org.apache.hadoop.hbase.HBaseClassTestRule;<a name="line.40"></a>
<span class="sourceLineNo">041</span>import org.apache.hadoop.hbase.HBaseTestingUtil;<a name="line.41"></a>
<span class="sourceLineNo">042</span>import org.apache.hadoop.hbase.HConstants;<a name="line.42"></a>
<span class="sourceLineNo">043</span>import org.apache.hadoop.hbase.TableName;<a name="line.43"></a>
<span class="sourceLineNo">044</span>import org.apache.hadoop.hbase.TableNotFoundException;<a name="line.44"></a>
<span class="sourceLineNo">045</span>import org.apache.hadoop.hbase.client.Connection;<a name="line.45"></a>
<span class="sourceLineNo">046</span>import org.apache.hadoop.hbase.client.ConnectionFactory;<a name="line.46"></a>
<span class="sourceLineNo">047</span>import org.apache.hadoop.hbase.client.Result;<a name="line.47"></a>
<span class="sourceLineNo">048</span>import org.apache.hadoop.hbase.client.ResultScanner;<a name="line.48"></a>
<span class="sourceLineNo">049</span>import org.apache.hadoop.hbase.client.Scan;<a name="line.49"></a>
<span class="sourceLineNo">050</span>import org.apache.hadoop.hbase.client.Table;<a name="line.50"></a>
<span class="sourceLineNo">051</span>import org.apache.hadoop.hbase.io.hfile.CacheConfig;<a name="line.51"></a>
<span class="sourceLineNo">052</span>import org.apache.hadoop.hbase.io.hfile.HFile;<a name="line.52"></a>
<span class="sourceLineNo">053</span>import org.apache.hadoop.hbase.io.hfile.HFileScanner;<a name="line.53"></a>
<span class="sourceLineNo">054</span>import org.apache.hadoop.hbase.regionserver.NoSuchColumnFamilyException;<a name="line.54"></a>
<span class="sourceLineNo">055</span>import org.apache.hadoop.hbase.testclassification.LargeTests;<a name="line.55"></a>
<span class="sourceLineNo">056</span>import org.apache.hadoop.hbase.testclassification.VerySlowMapReduceTests;<a name="line.56"></a>
<span class="sourceLineNo">057</span>import org.apache.hadoop.hbase.util.Bytes;<a name="line.57"></a>
<span class="sourceLineNo">058</span>import org.apache.hadoop.io.Text;<a name="line.58"></a>
<span class="sourceLineNo">059</span>import org.apache.hadoop.mapred.Utils.OutputFileUtils.OutputFilesFilter;<a name="line.59"></a>
<span class="sourceLineNo">060</span>import org.apache.hadoop.mapreduce.Job;<a name="line.60"></a>
<span class="sourceLineNo">061</span>import org.apache.hadoop.mapreduce.lib.output.NullOutputFormat;<a name="line.61"></a>
<span class="sourceLineNo">062</span>import org.apache.hadoop.util.Tool;<a name="line.62"></a>
<span class="sourceLineNo">063</span>import org.apache.hadoop.util.ToolRunner;<a name="line.63"></a>
<span class="sourceLineNo">064</span>import org.junit.AfterClass;<a name="line.64"></a>
<span class="sourceLineNo">065</span>import org.junit.Before;<a name="line.65"></a>
<span class="sourceLineNo">066</span>import org.junit.BeforeClass;<a name="line.66"></a>
<span class="sourceLineNo">067</span>import org.junit.ClassRule;<a name="line.67"></a>
<span class="sourceLineNo">068</span>import org.junit.Rule;<a name="line.68"></a>
<span class="sourceLineNo">069</span>import org.junit.Test;<a name="line.69"></a>
<span class="sourceLineNo">070</span>import org.junit.experimental.categories.Category;<a name="line.70"></a>
<span class="sourceLineNo">071</span>import org.junit.rules.ExpectedException;<a name="line.71"></a>
<span class="sourceLineNo">072</span>import org.slf4j.Logger;<a name="line.72"></a>
<span class="sourceLineNo">073</span>import org.slf4j.LoggerFactory;<a name="line.73"></a>
<span class="sourceLineNo">074</span><a name="line.74"></a>
<span class="sourceLineNo">075</span>@Category({ VerySlowMapReduceTests.class, LargeTests.class })<a name="line.75"></a>
<span class="sourceLineNo">076</span>public class TestImportTsv implements Configurable {<a name="line.76"></a>
<span class="sourceLineNo">077</span><a name="line.77"></a>
<span class="sourceLineNo">078</span> @ClassRule<a name="line.78"></a>
<span class="sourceLineNo">079</span> public static final HBaseClassTestRule CLASS_RULE =<a name="line.79"></a>
<span class="sourceLineNo">080</span> HBaseClassTestRule.forClass(TestImportTsv.class);<a name="line.80"></a>
<span class="sourceLineNo">081</span><a name="line.81"></a>
<span class="sourceLineNo">082</span> private static final Logger LOG = LoggerFactory.getLogger(TestImportTsv.class);<a name="line.82"></a>
<span class="sourceLineNo">083</span> protected static final String NAME = TestImportTsv.class.getSimpleName();<a name="line.83"></a>
<span class="sourceLineNo">084</span> protected static HBaseTestingUtil util = new HBaseTestingUtil();<a name="line.84"></a>
<span class="sourceLineNo">085</span><a name="line.85"></a>
<span class="sourceLineNo">086</span> // Delete the tmp directory after running doMROnTableTest. Boolean. Default is true.<a name="line.86"></a>
<span class="sourceLineNo">087</span> protected static final String DELETE_AFTER_LOAD_CONF = NAME + ".deleteAfterLoad";<a name="line.87"></a>
<span class="sourceLineNo">088</span><a name="line.88"></a>
<span class="sourceLineNo">089</span> /**<a name="line.89"></a>
<span class="sourceLineNo">090</span> * Force use of combiner in doMROnTableTest. Boolean. Default is true.<a name="line.90"></a>
<span class="sourceLineNo">091</span> */<a name="line.91"></a>
<span class="sourceLineNo">092</span> protected static final String FORCE_COMBINER_CONF = NAME + ".forceCombiner";<a name="line.92"></a>
<span class="sourceLineNo">093</span><a name="line.93"></a>
<span class="sourceLineNo">094</span> private final String FAMILY = "FAM";<a name="line.94"></a>
<span class="sourceLineNo">095</span> private TableName tn;<a name="line.95"></a>
<span class="sourceLineNo">096</span> private Map&lt;String, String&gt; args;<a name="line.96"></a>
<span class="sourceLineNo">097</span><a name="line.97"></a>
<span class="sourceLineNo">098</span> @Rule<a name="line.98"></a>
<span class="sourceLineNo">099</span> public ExpectedException exception = ExpectedException.none();<a name="line.99"></a>
<span class="sourceLineNo">100</span><a name="line.100"></a>
<span class="sourceLineNo">101</span> public Configuration getConf() {<a name="line.101"></a>
<span class="sourceLineNo">102</span> return util.getConfiguration();<a name="line.102"></a>
<span class="sourceLineNo">103</span> }<a name="line.103"></a>
<span class="sourceLineNo">104</span><a name="line.104"></a>
<span class="sourceLineNo">105</span> public void setConf(Configuration conf) {<a name="line.105"></a>
<span class="sourceLineNo">106</span> throw new IllegalArgumentException("setConf not supported");<a name="line.106"></a>
<span class="sourceLineNo">107</span> }<a name="line.107"></a>
<span class="sourceLineNo">108</span><a name="line.108"></a>
<span class="sourceLineNo">109</span> @BeforeClass<a name="line.109"></a>
<span class="sourceLineNo">110</span> public static void provisionCluster() throws Exception {<a name="line.110"></a>
<span class="sourceLineNo">111</span> util.startMiniCluster();<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> @AfterClass<a name="line.114"></a>
<span class="sourceLineNo">115</span> public static void releaseCluster() throws Exception {<a name="line.115"></a>
<span class="sourceLineNo">116</span> util.shutdownMiniCluster();<a name="line.116"></a>
<span class="sourceLineNo">117</span> }<a name="line.117"></a>
<span class="sourceLineNo">118</span><a name="line.118"></a>
<span class="sourceLineNo">119</span> @Before<a name="line.119"></a>
<span class="sourceLineNo">120</span> public void setup() throws Exception {<a name="line.120"></a>
<span class="sourceLineNo">121</span> tn = TableName.valueOf("test-" + util.getRandomUUID());<a name="line.121"></a>
<span class="sourceLineNo">122</span> args = new HashMap&lt;&gt;();<a name="line.122"></a>
<span class="sourceLineNo">123</span> // Prepare the arguments required for the test.<a name="line.123"></a>
<span class="sourceLineNo">124</span> args.put(ImportTsv.COLUMNS_CONF_KEY, "HBASE_ROW_KEY,FAM:A,FAM:B");<a name="line.124"></a>
<span class="sourceLineNo">125</span> args.put(ImportTsv.SEPARATOR_CONF_KEY, "\u001b");<a name="line.125"></a>
<span class="sourceLineNo">126</span> }<a name="line.126"></a>
<span class="sourceLineNo">127</span><a name="line.127"></a>
<span class="sourceLineNo">128</span> @Test<a name="line.128"></a>
<span class="sourceLineNo">129</span> public void testMROnTable() throws Exception {<a name="line.129"></a>
<span class="sourceLineNo">130</span> util.createTable(tn, FAMILY);<a name="line.130"></a>
<span class="sourceLineNo">131</span> doMROnTableTest(null, 1);<a name="line.131"></a>
<span class="sourceLineNo">132</span> util.deleteTable(tn);<a name="line.132"></a>
<span class="sourceLineNo">133</span> }<a name="line.133"></a>
<span class="sourceLineNo">134</span><a name="line.134"></a>
<span class="sourceLineNo">135</span> @Test<a name="line.135"></a>
<span class="sourceLineNo">136</span> public void testMROnTableWithTimestamp() throws Exception {<a name="line.136"></a>
<span class="sourceLineNo">137</span> util.createTable(tn, FAMILY);<a name="line.137"></a>
<span class="sourceLineNo">138</span> args.put(ImportTsv.COLUMNS_CONF_KEY, "HBASE_ROW_KEY,HBASE_TS_KEY,FAM:A,FAM:B");<a name="line.138"></a>
<span class="sourceLineNo">139</span> args.put(ImportTsv.SEPARATOR_CONF_KEY, ",");<a name="line.139"></a>
<span class="sourceLineNo">140</span> String data = "KEY,1234,VALUE1,VALUE2\n";<a name="line.140"></a>
<span class="sourceLineNo">141</span><a name="line.141"></a>
<span class="sourceLineNo">142</span> doMROnTableTest(data, 1);<a name="line.142"></a>
<span class="sourceLineNo">143</span> util.deleteTable(tn);<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> @Test<a name="line.146"></a>
<span class="sourceLineNo">147</span> public void testMROnTableWithCustomMapper() throws Exception {<a name="line.147"></a>
<span class="sourceLineNo">148</span> util.createTable(tn, FAMILY);<a name="line.148"></a>
<span class="sourceLineNo">149</span> args.put(ImportTsv.MAPPER_CONF_KEY,<a name="line.149"></a>
<span class="sourceLineNo">150</span> "org.apache.hadoop.hbase.mapreduce.TsvImporterCustomTestMapper");<a name="line.150"></a>
<span class="sourceLineNo">151</span><a name="line.151"></a>
<span class="sourceLineNo">152</span> doMROnTableTest(null, 3);<a name="line.152"></a>
<span class="sourceLineNo">153</span> util.deleteTable(tn);<a name="line.153"></a>
<span class="sourceLineNo">154</span> }<a name="line.154"></a>
<span class="sourceLineNo">155</span><a name="line.155"></a>
<span class="sourceLineNo">156</span> @Test<a name="line.156"></a>
<span class="sourceLineNo">157</span> public void testBulkOutputWithoutAnExistingTable() throws Exception {<a name="line.157"></a>
<span class="sourceLineNo">158</span> // Prepare the arguments required for the test.<a name="line.158"></a>
<span class="sourceLineNo">159</span> Path hfiles = new Path(util.getDataTestDirOnTestFS(tn.getNameAsString()), "hfiles");<a name="line.159"></a>
<span class="sourceLineNo">160</span> args.put(ImportTsv.BULK_OUTPUT_CONF_KEY, hfiles.toString());<a name="line.160"></a>
<span class="sourceLineNo">161</span><a name="line.161"></a>
<span class="sourceLineNo">162</span> doMROnTableTest(null, 3);<a name="line.162"></a>
<span class="sourceLineNo">163</span> util.deleteTable(tn);<a name="line.163"></a>
<span class="sourceLineNo">164</span> }<a name="line.164"></a>
<span class="sourceLineNo">165</span><a name="line.165"></a>
<span class="sourceLineNo">166</span> @Test<a name="line.166"></a>
<span class="sourceLineNo">167</span> public void testBulkOutputWithAnExistingTable() throws Exception {<a name="line.167"></a>
<span class="sourceLineNo">168</span> util.createTable(tn, FAMILY);<a name="line.168"></a>
<span class="sourceLineNo">169</span><a name="line.169"></a>
<span class="sourceLineNo">170</span> // Prepare the arguments required for the test.<a name="line.170"></a>
<span class="sourceLineNo">171</span> Path hfiles = new Path(util.getDataTestDirOnTestFS(tn.getNameAsString()), "hfiles");<a name="line.171"></a>
<span class="sourceLineNo">172</span> args.put(ImportTsv.BULK_OUTPUT_CONF_KEY, hfiles.toString());<a name="line.172"></a>
<span class="sourceLineNo">173</span><a name="line.173"></a>
<span class="sourceLineNo">174</span> doMROnTableTest(null, 3);<a name="line.174"></a>
<span class="sourceLineNo">175</span> util.deleteTable(tn);<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> @Test<a name="line.178"></a>
<span class="sourceLineNo">179</span> public void testBulkOutputWithAnExistingTableNoStrictTrue() throws Exception {<a name="line.179"></a>
<span class="sourceLineNo">180</span> util.createTable(tn, FAMILY);<a name="line.180"></a>
<span class="sourceLineNo">181</span><a name="line.181"></a>
<span class="sourceLineNo">182</span> // Prepare the arguments required for the test.<a name="line.182"></a>
<span class="sourceLineNo">183</span> Path hfiles = new Path(util.getDataTestDirOnTestFS(tn.getNameAsString()), "hfiles");<a name="line.183"></a>
<span class="sourceLineNo">184</span> args.put(ImportTsv.BULK_OUTPUT_CONF_KEY, hfiles.toString());<a name="line.184"></a>
<span class="sourceLineNo">185</span> args.put(ImportTsv.NO_STRICT_COL_FAMILY, "true");<a name="line.185"></a>
<span class="sourceLineNo">186</span> doMROnTableTest(null, 3);<a name="line.186"></a>
<span class="sourceLineNo">187</span> util.deleteTable(tn);<a name="line.187"></a>
<span class="sourceLineNo">188</span> }<a name="line.188"></a>
<span class="sourceLineNo">189</span><a name="line.189"></a>
<span class="sourceLineNo">190</span> @Test<a name="line.190"></a>
<span class="sourceLineNo">191</span> public void testJobConfigurationsWithTsvImporterTextMapper() throws Exception {<a name="line.191"></a>
<span class="sourceLineNo">192</span> Path bulkOutputPath = new Path(util.getDataTestDirOnTestFS(tn.getNameAsString()), "hfiles");<a name="line.192"></a>
<span class="sourceLineNo">193</span> String INPUT_FILE = "InputFile1.csv";<a name="line.193"></a>
<span class="sourceLineNo">194</span> // Prepare the arguments required for the test.<a name="line.194"></a>
<span class="sourceLineNo">195</span> String[] args = new String[] {<a name="line.195"></a>
<span class="sourceLineNo">196</span> "-D" + ImportTsv.MAPPER_CONF_KEY + "=org.apache.hadoop.hbase.mapreduce.TsvImporterTextMapper",<a name="line.196"></a>
<span class="sourceLineNo">197</span> "-D" + ImportTsv.COLUMNS_CONF_KEY + "=HBASE_ROW_KEY,FAM:A,FAM:B",<a name="line.197"></a>
<span class="sourceLineNo">198</span> "-D" + ImportTsv.SEPARATOR_CONF_KEY + "=,",<a name="line.198"></a>
<span class="sourceLineNo">199</span> "-D" + ImportTsv.BULK_OUTPUT_CONF_KEY + "=" + bulkOutputPath.toString(), tn.getNameAsString(),<a name="line.199"></a>
<span class="sourceLineNo">200</span> INPUT_FILE };<a name="line.200"></a>
<span class="sourceLineNo">201</span> assertEquals("running test job configuration failed.", 0,<a name="line.201"></a>
<span class="sourceLineNo">202</span> ToolRunner.run(new Configuration(util.getConfiguration()), new ImportTsv() {<a name="line.202"></a>
<span class="sourceLineNo">203</span> @Override<a name="line.203"></a>
<span class="sourceLineNo">204</span> public int run(String[] args) throws Exception {<a name="line.204"></a>
<span class="sourceLineNo">205</span> Job job = createSubmittableJob(getConf(), args);<a name="line.205"></a>
<span class="sourceLineNo">206</span> assertTrue(job.getMapperClass().equals(TsvImporterTextMapper.class));<a name="line.206"></a>
<span class="sourceLineNo">207</span> assertTrue(job.getReducerClass().equals(TextSortReducer.class));<a name="line.207"></a>
<span class="sourceLineNo">208</span> assertTrue(job.getMapOutputValueClass().equals(Text.class));<a name="line.208"></a>
<span class="sourceLineNo">209</span> return 0;<a name="line.209"></a>
<span class="sourceLineNo">210</span> }<a name="line.210"></a>
<span class="sourceLineNo">211</span> }, args));<a name="line.211"></a>
<span class="sourceLineNo">212</span> // Delete table created by createSubmittableJob.<a name="line.212"></a>
<span class="sourceLineNo">213</span> util.deleteTable(tn);<a name="line.213"></a>
<span class="sourceLineNo">214</span> }<a name="line.214"></a>
<span class="sourceLineNo">215</span><a name="line.215"></a>
<span class="sourceLineNo">216</span> @Test<a name="line.216"></a>
<span class="sourceLineNo">217</span> public void testBulkOutputWithTsvImporterTextMapper() throws Exception {<a name="line.217"></a>
<span class="sourceLineNo">218</span> Path bulkOutputPath = new Path(util.getDataTestDirOnTestFS(tn.getNameAsString()), "hfiles");<a name="line.218"></a>
<span class="sourceLineNo">219</span> args.put(ImportTsv.MAPPER_CONF_KEY, "org.apache.hadoop.hbase.mapreduce.TsvImporterTextMapper");<a name="line.219"></a>
<span class="sourceLineNo">220</span> args.put(ImportTsv.BULK_OUTPUT_CONF_KEY, bulkOutputPath.toString());<a name="line.220"></a>
<span class="sourceLineNo">221</span> String data = "KEY\u001bVALUE4\u001bVALUE8\n";<a name="line.221"></a>
<span class="sourceLineNo">222</span> doMROnTableTest(data, 4);<a name="line.222"></a>
<span class="sourceLineNo">223</span> util.deleteTable(tn);<a name="line.223"></a>
<span class="sourceLineNo">224</span> }<a name="line.224"></a>
<span class="sourceLineNo">225</span><a name="line.225"></a>
<span class="sourceLineNo">226</span> @Test<a name="line.226"></a>
<span class="sourceLineNo">227</span> public void testWithoutAnExistingTableAndCreateTableSetToNo() throws Exception {<a name="line.227"></a>
<span class="sourceLineNo">228</span> String[] args = new String[] { tn.getNameAsString(), "/inputFile" };<a name="line.228"></a>
<span class="sourceLineNo">229</span><a name="line.229"></a>
<span class="sourceLineNo">230</span> Configuration conf = new Configuration(util.getConfiguration());<a name="line.230"></a>
<span class="sourceLineNo">231</span> conf.set(ImportTsv.COLUMNS_CONF_KEY, "HBASE_ROW_KEY,FAM:A");<a name="line.231"></a>
<span class="sourceLineNo">232</span> conf.set(ImportTsv.BULK_OUTPUT_CONF_KEY, "/output");<a name="line.232"></a>
<span class="sourceLineNo">233</span> conf.set(ImportTsv.CREATE_TABLE_CONF_KEY, "no");<a name="line.233"></a>
<span class="sourceLineNo">234</span> exception.expect(TableNotFoundException.class);<a name="line.234"></a>
<span class="sourceLineNo">235</span> assertEquals("running test job configuration failed.", 0,<a name="line.235"></a>
<span class="sourceLineNo">236</span> ToolRunner.run(new Configuration(util.getConfiguration()), new ImportTsv() {<a name="line.236"></a>
<span class="sourceLineNo">237</span> @Override<a name="line.237"></a>
<span class="sourceLineNo">238</span> public int run(String[] args) throws Exception {<a name="line.238"></a>
<span class="sourceLineNo">239</span> createSubmittableJob(getConf(), args);<a name="line.239"></a>
<span class="sourceLineNo">240</span> return 0;<a name="line.240"></a>
<span class="sourceLineNo">241</span> }<a name="line.241"></a>
<span class="sourceLineNo">242</span> }, args));<a name="line.242"></a>
<span class="sourceLineNo">243</span> }<a name="line.243"></a>
<span class="sourceLineNo">244</span><a name="line.244"></a>
<span class="sourceLineNo">245</span> @Test<a name="line.245"></a>
<span class="sourceLineNo">246</span> public void testMRNoMatchedColumnFamily() throws Exception {<a name="line.246"></a>
<span class="sourceLineNo">247</span> util.createTable(tn, FAMILY);<a name="line.247"></a>
<span class="sourceLineNo">248</span><a name="line.248"></a>
<span class="sourceLineNo">249</span> String[] args = new String[] {<a name="line.249"></a>
<span class="sourceLineNo">250</span> "-D" + ImportTsv.COLUMNS_CONF_KEY<a name="line.250"></a>
<span class="sourceLineNo">251</span> + "=HBASE_ROW_KEY,FAM:A,FAM01_ERROR:A,FAM01_ERROR:B,FAM02_ERROR:C",<a name="line.251"></a>
<span class="sourceLineNo">252</span> tn.getNameAsString(), "/inputFile" };<a name="line.252"></a>
<span class="sourceLineNo">253</span> exception.expect(NoSuchColumnFamilyException.class);<a name="line.253"></a>
<span class="sourceLineNo">254</span> assertEquals("running test job configuration failed.", 0,<a name="line.254"></a>
<span class="sourceLineNo">255</span> ToolRunner.run(new Configuration(util.getConfiguration()), new ImportTsv() {<a name="line.255"></a>
<span class="sourceLineNo">256</span> @Override<a name="line.256"></a>
<span class="sourceLineNo">257</span> public int run(String[] args) throws Exception {<a name="line.257"></a>
<span class="sourceLineNo">258</span> createSubmittableJob(getConf(), args);<a name="line.258"></a>
<span class="sourceLineNo">259</span> return 0;<a name="line.259"></a>
<span class="sourceLineNo">260</span> }<a name="line.260"></a>
<span class="sourceLineNo">261</span> }, args));<a name="line.261"></a>
<span class="sourceLineNo">262</span><a name="line.262"></a>
<span class="sourceLineNo">263</span> util.deleteTable(tn);<a name="line.263"></a>
<span class="sourceLineNo">264</span> }<a name="line.264"></a>
<span class="sourceLineNo">265</span><a name="line.265"></a>
<span class="sourceLineNo">266</span> @Test<a name="line.266"></a>
<span class="sourceLineNo">267</span> public void testMRWithoutAnExistingTable() throws Exception {<a name="line.267"></a>
<span class="sourceLineNo">268</span> String[] args = new String[] { tn.getNameAsString(), "/inputFile" };<a name="line.268"></a>
<span class="sourceLineNo">269</span><a name="line.269"></a>
<span class="sourceLineNo">270</span> exception.expect(TableNotFoundException.class);<a name="line.270"></a>
<span class="sourceLineNo">271</span> assertEquals("running test job configuration failed.", 0,<a name="line.271"></a>
<span class="sourceLineNo">272</span> ToolRunner.run(new Configuration(util.getConfiguration()), new ImportTsv() {<a name="line.272"></a>
<span class="sourceLineNo">273</span> @Override<a name="line.273"></a>
<span class="sourceLineNo">274</span> public int run(String[] args) throws Exception {<a name="line.274"></a>
<span class="sourceLineNo">275</span> createSubmittableJob(getConf(), args);<a name="line.275"></a>
<span class="sourceLineNo">276</span> return 0;<a name="line.276"></a>
<span class="sourceLineNo">277</span> }<a name="line.277"></a>
<span class="sourceLineNo">278</span> }, args));<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> @Test<a name="line.281"></a>
<span class="sourceLineNo">282</span> public void testJobConfigurationsWithDryMode() throws Exception {<a name="line.282"></a>
<span class="sourceLineNo">283</span> Path bulkOutputPath = new Path(util.getDataTestDirOnTestFS(tn.getNameAsString()), "hfiles");<a name="line.283"></a>
<span class="sourceLineNo">284</span> String INPUT_FILE = "InputFile1.csv";<a name="line.284"></a>
<span class="sourceLineNo">285</span> // Prepare the arguments required for the test.<a name="line.285"></a>
<span class="sourceLineNo">286</span> String[] argsArray =<a name="line.286"></a>
<span class="sourceLineNo">287</span> new String[] { "-D" + ImportTsv.COLUMNS_CONF_KEY + "=HBASE_ROW_KEY,FAM:A,FAM:B",<a name="line.287"></a>
<span class="sourceLineNo">288</span> "-D" + ImportTsv.SEPARATOR_CONF_KEY + "=,",<a name="line.288"></a>
<span class="sourceLineNo">289</span> "-D" + ImportTsv.BULK_OUTPUT_CONF_KEY + "=" + bulkOutputPath.toString(),<a name="line.289"></a>
<span class="sourceLineNo">290</span> "-D" + ImportTsv.DRY_RUN_CONF_KEY + "=true", tn.getNameAsString(), INPUT_FILE };<a name="line.290"></a>
<span class="sourceLineNo">291</span> assertEquals("running test job configuration failed.", 0,<a name="line.291"></a>
<span class="sourceLineNo">292</span> ToolRunner.run(new Configuration(util.getConfiguration()), new ImportTsv() {<a name="line.292"></a>
<span class="sourceLineNo">293</span> @Override<a name="line.293"></a>
<span class="sourceLineNo">294</span> public int run(String[] args) throws Exception {<a name="line.294"></a>
<span class="sourceLineNo">295</span> Job job = createSubmittableJob(getConf(), args);<a name="line.295"></a>
<span class="sourceLineNo">296</span> assertTrue(job.getOutputFormatClass().equals(NullOutputFormat.class));<a name="line.296"></a>
<span class="sourceLineNo">297</span> return 0;<a name="line.297"></a>
<span class="sourceLineNo">298</span> }<a name="line.298"></a>
<span class="sourceLineNo">299</span> }, argsArray));<a name="line.299"></a>
<span class="sourceLineNo">300</span> // Delete table created by createSubmittableJob.<a name="line.300"></a>
<span class="sourceLineNo">301</span> util.deleteTable(tn);<a name="line.301"></a>
<span class="sourceLineNo">302</span> }<a name="line.302"></a>
<span class="sourceLineNo">303</span><a name="line.303"></a>
<span class="sourceLineNo">304</span> @Test<a name="line.304"></a>
<span class="sourceLineNo">305</span> public void testDryModeWithoutBulkOutputAndTableExists() throws Exception {<a name="line.305"></a>
<span class="sourceLineNo">306</span> util.createTable(tn, FAMILY);<a name="line.306"></a>
<span class="sourceLineNo">307</span> args.put(ImportTsv.DRY_RUN_CONF_KEY, "true");<a name="line.307"></a>
<span class="sourceLineNo">308</span> doMROnTableTest(null, 1);<a name="line.308"></a>
<span class="sourceLineNo">309</span> // Dry mode should not delete an existing table. If it's not present,<a name="line.309"></a>
<span class="sourceLineNo">310</span> // this will throw TableNotFoundException.<a name="line.310"></a>
<span class="sourceLineNo">311</span> util.deleteTable(tn);<a name="line.311"></a>
<span class="sourceLineNo">312</span> }<a name="line.312"></a>
<span class="sourceLineNo">313</span><a name="line.313"></a>
<span class="sourceLineNo">314</span> /**<a name="line.314"></a>
<span class="sourceLineNo">315</span> * If table is not present in non-bulk mode, dry run should fail just like normal mode.<a name="line.315"></a>
<span class="sourceLineNo">316</span> */<a name="line.316"></a>
<span class="sourceLineNo">317</span> @Test<a name="line.317"></a>
<span class="sourceLineNo">318</span> public void testDryModeWithoutBulkOutputAndTableDoesNotExists() throws Exception {<a name="line.318"></a>
<span class="sourceLineNo">319</span> args.put(ImportTsv.DRY_RUN_CONF_KEY, "true");<a name="line.319"></a>
<span class="sourceLineNo">320</span> exception.expect(TableNotFoundException.class);<a name="line.320"></a>
<span class="sourceLineNo">321</span> doMROnTableTest(null, 1);<a name="line.321"></a>
<span class="sourceLineNo">322</span> }<a name="line.322"></a>
<span class="sourceLineNo">323</span><a name="line.323"></a>
<span class="sourceLineNo">324</span> @Test<a name="line.324"></a>
<span class="sourceLineNo">325</span> public void testDryModeWithBulkOutputAndTableExists() throws Exception {<a name="line.325"></a>
<span class="sourceLineNo">326</span> util.createTable(tn, FAMILY);<a name="line.326"></a>
<span class="sourceLineNo">327</span> // Prepare the arguments required for the test.<a name="line.327"></a>
<span class="sourceLineNo">328</span> Path hfiles = new Path(util.getDataTestDirOnTestFS(tn.getNameAsString()), "hfiles");<a name="line.328"></a>
<span class="sourceLineNo">329</span> args.put(ImportTsv.BULK_OUTPUT_CONF_KEY, hfiles.toString());<a name="line.329"></a>
<span class="sourceLineNo">330</span> args.put(ImportTsv.DRY_RUN_CONF_KEY, "true");<a name="line.330"></a>
<span class="sourceLineNo">331</span> doMROnTableTest(null, 1);<a name="line.331"></a>
<span class="sourceLineNo">332</span> // Dry mode should not delete an existing table. If it's not present,<a name="line.332"></a>
<span class="sourceLineNo">333</span> // this will throw TableNotFoundException.<a name="line.333"></a>
<span class="sourceLineNo">334</span> util.deleteTable(tn);<a name="line.334"></a>
<span class="sourceLineNo">335</span> }<a name="line.335"></a>
<span class="sourceLineNo">336</span><a name="line.336"></a>
<span class="sourceLineNo">337</span> /**<a name="line.337"></a>
<span class="sourceLineNo">338</span> * If table is not present in bulk mode and create.table is not set to yes, import should fail<a name="line.338"></a>
<span class="sourceLineNo">339</span> * with TableNotFoundException.<a name="line.339"></a>
<span class="sourceLineNo">340</span> */<a name="line.340"></a>
<span class="sourceLineNo">341</span> @Test<a name="line.341"></a>
<span class="sourceLineNo">342</span> public void testDryModeWithBulkOutputAndTableDoesNotExistsCreateTableSetToNo() throws Exception {<a name="line.342"></a>
<span class="sourceLineNo">343</span> // Prepare the arguments required for the test.<a name="line.343"></a>
<span class="sourceLineNo">344</span> Path hfiles = new Path(util.getDataTestDirOnTestFS(tn.getNameAsString()), "hfiles");<a name="line.344"></a>
<span class="sourceLineNo">345</span> args.put(ImportTsv.BULK_OUTPUT_CONF_KEY, hfiles.toString());<a name="line.345"></a>
<span class="sourceLineNo">346</span> args.put(ImportTsv.DRY_RUN_CONF_KEY, "true");<a name="line.346"></a>
<span class="sourceLineNo">347</span> args.put(ImportTsv.CREATE_TABLE_CONF_KEY, "no");<a name="line.347"></a>
<span class="sourceLineNo">348</span> exception.expect(TableNotFoundException.class);<a name="line.348"></a>
<span class="sourceLineNo">349</span> doMROnTableTest(null, 1);<a name="line.349"></a>
<span class="sourceLineNo">350</span> }<a name="line.350"></a>
<span class="sourceLineNo">351</span><a name="line.351"></a>
<span class="sourceLineNo">352</span> @Test<a name="line.352"></a>
<span class="sourceLineNo">353</span> public void testDryModeWithBulkModeAndTableDoesNotExistsCreateTableSetToYes() throws Exception {<a name="line.353"></a>
<span class="sourceLineNo">354</span> // Prepare the arguments required for the test.<a name="line.354"></a>
<span class="sourceLineNo">355</span> Path hfiles = new Path(util.getDataTestDirOnTestFS(tn.getNameAsString()), "hfiles");<a name="line.355"></a>
<span class="sourceLineNo">356</span> args.put(ImportTsv.BULK_OUTPUT_CONF_KEY, hfiles.toString());<a name="line.356"></a>
<span class="sourceLineNo">357</span> args.put(ImportTsv.DRY_RUN_CONF_KEY, "true");<a name="line.357"></a>
<span class="sourceLineNo">358</span> args.put(ImportTsv.CREATE_TABLE_CONF_KEY, "yes");<a name="line.358"></a>
<span class="sourceLineNo">359</span> doMROnTableTest(null, 1);<a name="line.359"></a>
<span class="sourceLineNo">360</span> // Verify temporary table was deleted.<a name="line.360"></a>
<span class="sourceLineNo">361</span> exception.expect(TableNotFoundException.class);<a name="line.361"></a>
<span class="sourceLineNo">362</span> util.deleteTable(tn);<a name="line.362"></a>
<span class="sourceLineNo">363</span> }<a name="line.363"></a>
<span class="sourceLineNo">364</span><a name="line.364"></a>
<span class="sourceLineNo">365</span> /**<a name="line.365"></a>
<span class="sourceLineNo">366</span> * If there are invalid data rows as inputs, then only those rows should be ignored.<a name="line.366"></a>
<span class="sourceLineNo">367</span> */<a name="line.367"></a>
<span class="sourceLineNo">368</span> @Test<a name="line.368"></a>
<span class="sourceLineNo">369</span> public void testTsvImporterTextMapperWithInvalidData() throws Exception {<a name="line.369"></a>
<span class="sourceLineNo">370</span> Path bulkOutputPath = new Path(util.getDataTestDirOnTestFS(tn.getNameAsString()), "hfiles");<a name="line.370"></a>
<span class="sourceLineNo">371</span> args.put(ImportTsv.MAPPER_CONF_KEY, "org.apache.hadoop.hbase.mapreduce.TsvImporterTextMapper");<a name="line.371"></a>
<span class="sourceLineNo">372</span> args.put(ImportTsv.BULK_OUTPUT_CONF_KEY, bulkOutputPath.toString());<a name="line.372"></a>
<span class="sourceLineNo">373</span> args.put(ImportTsv.COLUMNS_CONF_KEY, "HBASE_ROW_KEY,HBASE_TS_KEY,FAM:A,FAM:B");<a name="line.373"></a>
<span class="sourceLineNo">374</span> args.put(ImportTsv.SEPARATOR_CONF_KEY, ",");<a name="line.374"></a>
<span class="sourceLineNo">375</span> // 3 Rows of data as input. 2 Rows are valid and 1 row is invalid as it doesn't have TS<a name="line.375"></a>
<span class="sourceLineNo">376</span> String data = "KEY,1234,VALUE1,VALUE2\nKEY\nKEY,1235,VALUE1,VALUE2\n";<a name="line.376"></a>
<span class="sourceLineNo">377</span> doMROnTableTest(util, tn, FAMILY, data, args, 1, 4);<a name="line.377"></a>
<span class="sourceLineNo">378</span> util.deleteTable(tn);<a name="line.378"></a>
<span class="sourceLineNo">379</span> }<a name="line.379"></a>
<span class="sourceLineNo">380</span><a name="line.380"></a>
<span class="sourceLineNo">381</span> @Test<a name="line.381"></a>
<span class="sourceLineNo">382</span> public void testSkipEmptyColumns() throws Exception {<a name="line.382"></a>
<span class="sourceLineNo">383</span> Path bulkOutputPath = new Path(util.getDataTestDirOnTestFS(tn.getNameAsString()), "hfiles");<a name="line.383"></a>
<span class="sourceLineNo">384</span> args.put(ImportTsv.BULK_OUTPUT_CONF_KEY, bulkOutputPath.toString());<a name="line.384"></a>
<span class="sourceLineNo">385</span> args.put(ImportTsv.COLUMNS_CONF_KEY, "HBASE_ROW_KEY,HBASE_TS_KEY,FAM:A,FAM:B");<a name="line.385"></a>
<span class="sourceLineNo">386</span> args.put(ImportTsv.SEPARATOR_CONF_KEY, ",");<a name="line.386"></a>
<span class="sourceLineNo">387</span> args.put(ImportTsv.SKIP_EMPTY_COLUMNS, "true");<a name="line.387"></a>
<span class="sourceLineNo">388</span> // 2 Rows of data as input. Both rows are valid and only 3 columns are no-empty among 4<a name="line.388"></a>
<span class="sourceLineNo">389</span> String data = "KEY,1234,VALUE1,VALUE2\nKEY,1235,,VALUE2\n";<a name="line.389"></a>
<span class="sourceLineNo">390</span> doMROnTableTest(util, tn, FAMILY, data, args, 1, 3);<a name="line.390"></a>
<span class="sourceLineNo">391</span> util.deleteTable(tn);<a name="line.391"></a>
<span class="sourceLineNo">392</span> }<a name="line.392"></a>
<span class="sourceLineNo">393</span><a name="line.393"></a>
<span class="sourceLineNo">394</span> private Tool doMROnTableTest(String data, int valueMultiplier) throws Exception {<a name="line.394"></a>
<span class="sourceLineNo">395</span> return doMROnTableTest(util, tn, FAMILY, data, args, valueMultiplier, -1);<a name="line.395"></a>
<span class="sourceLineNo">396</span> }<a name="line.396"></a>
<span class="sourceLineNo">397</span><a name="line.397"></a>
<span class="sourceLineNo">398</span> protected static Tool doMROnTableTest(HBaseTestingUtil util, TableName table, String family,<a name="line.398"></a>
<span class="sourceLineNo">399</span> String data, Map&lt;String, String&gt; args) throws Exception {<a name="line.399"></a>
<span class="sourceLineNo">400</span> return doMROnTableTest(util, table, family, data, args, 1, -1);<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> /**<a name="line.403"></a>
<span class="sourceLineNo">404</span> * Run an ImportTsv job and perform basic validation on the results. Returns the ImportTsv<a name="line.404"></a>
<span class="sourceLineNo">405</span> * &lt;code&gt;Tool&lt;/code&gt; instance so that other tests can inspect it for further validation as<a name="line.405"></a>
<span class="sourceLineNo">406</span> * necessary. This method is static to insure non-reliance on instance's util/conf facilities.<a name="line.406"></a>
<span class="sourceLineNo">407</span> * @param args Any arguments to pass BEFORE inputFile path is appended.<a name="line.407"></a>
<span class="sourceLineNo">408</span> * @return The Tool instance used to run the test.<a name="line.408"></a>
<span class="sourceLineNo">409</span> */<a name="line.409"></a>
<span class="sourceLineNo">410</span> protected static Tool doMROnTableTest(HBaseTestingUtil util, TableName table, String family,<a name="line.410"></a>
<span class="sourceLineNo">411</span> String data, Map&lt;String, String&gt; args, int valueMultiplier, int expectedKVCount)<a name="line.411"></a>
<span class="sourceLineNo">412</span> throws Exception {<a name="line.412"></a>
<span class="sourceLineNo">413</span> Configuration conf = new Configuration(util.getConfiguration());<a name="line.413"></a>
<span class="sourceLineNo">414</span><a name="line.414"></a>
<span class="sourceLineNo">415</span> // populate input file<a name="line.415"></a>
<span class="sourceLineNo">416</span> FileSystem fs = FileSystem.get(conf);<a name="line.416"></a>
<span class="sourceLineNo">417</span> Path inputPath =<a name="line.417"></a>
<span class="sourceLineNo">418</span> fs.makeQualified(new Path(util.getDataTestDirOnTestFS(table.getNameAsString()), "input.dat"));<a name="line.418"></a>
<span class="sourceLineNo">419</span> FSDataOutputStream op = fs.create(inputPath, true);<a name="line.419"></a>
<span class="sourceLineNo">420</span> if (data == null) {<a name="line.420"></a>
<span class="sourceLineNo">421</span> data = "KEY\u001bVALUE1\u001bVALUE2\n";<a name="line.421"></a>
<span class="sourceLineNo">422</span> }<a name="line.422"></a>
<span class="sourceLineNo">423</span> op.write(Bytes.toBytes(data));<a name="line.423"></a>
<span class="sourceLineNo">424</span> op.close();<a name="line.424"></a>
<span class="sourceLineNo">425</span> LOG.debug(String.format("Wrote test data to file: %s", inputPath));<a name="line.425"></a>
<span class="sourceLineNo">426</span><a name="line.426"></a>
<span class="sourceLineNo">427</span> if (conf.getBoolean(FORCE_COMBINER_CONF, true)) {<a name="line.427"></a>
<span class="sourceLineNo">428</span> LOG.debug("Forcing combiner.");<a name="line.428"></a>
<span class="sourceLineNo">429</span> conf.setInt("mapreduce.map.combine.minspills", 1);<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> // Build args array.<a name="line.432"></a>
<span class="sourceLineNo">433</span> String[] argsArray = new String[args.size() + 2];<a name="line.433"></a>
<span class="sourceLineNo">434</span> Iterator it = args.entrySet().iterator();<a name="line.434"></a>
<span class="sourceLineNo">435</span> int i = 0;<a name="line.435"></a>
<span class="sourceLineNo">436</span> while (it.hasNext()) {<a name="line.436"></a>
<span class="sourceLineNo">437</span> Map.Entry pair = (Map.Entry) it.next();<a name="line.437"></a>
<span class="sourceLineNo">438</span> argsArray[i] = "-D" + pair.getKey() + "=" + pair.getValue();<a name="line.438"></a>
<span class="sourceLineNo">439</span> i++;<a name="line.439"></a>
<span class="sourceLineNo">440</span> }<a name="line.440"></a>
<span class="sourceLineNo">441</span> argsArray[i] = table.getNameAsString();<a name="line.441"></a>
<span class="sourceLineNo">442</span> argsArray[i + 1] = inputPath.toString();<a name="line.442"></a>
<span class="sourceLineNo">443</span><a name="line.443"></a>
<span class="sourceLineNo">444</span> // run the import<a name="line.444"></a>
<span class="sourceLineNo">445</span> Tool tool = new ImportTsv();<a name="line.445"></a>
<span class="sourceLineNo">446</span> LOG.debug("Running ImportTsv with arguments: " + Arrays.toString(argsArray));<a name="line.446"></a>
<span class="sourceLineNo">447</span> assertEquals(0, ToolRunner.run(conf, tool, argsArray));<a name="line.447"></a>
<span class="sourceLineNo">448</span><a name="line.448"></a>
<span class="sourceLineNo">449</span> // Perform basic validation. If the input args did not include<a name="line.449"></a>
<span class="sourceLineNo">450</span> // ImportTsv.BULK_OUTPUT_CONF_KEY then validate data in the table.<a name="line.450"></a>
<span class="sourceLineNo">451</span> // Otherwise, validate presence of hfiles.<a name="line.451"></a>
<span class="sourceLineNo">452</span> boolean isDryRun = args.containsKey(ImportTsv.DRY_RUN_CONF_KEY)<a name="line.452"></a>
<span class="sourceLineNo">453</span> &amp;&amp; "true".equalsIgnoreCase(args.get(ImportTsv.DRY_RUN_CONF_KEY));<a name="line.453"></a>
<span class="sourceLineNo">454</span> if (args.containsKey(ImportTsv.BULK_OUTPUT_CONF_KEY)) {<a name="line.454"></a>
<span class="sourceLineNo">455</span> if (isDryRun) {<a name="line.455"></a>
<span class="sourceLineNo">456</span> assertFalse(String.format("Dry run mode, %s should not have been created.",<a name="line.456"></a>
<span class="sourceLineNo">457</span> ImportTsv.BULK_OUTPUT_CONF_KEY), fs.exists(new Path(ImportTsv.BULK_OUTPUT_CONF_KEY)));<a name="line.457"></a>
<span class="sourceLineNo">458</span> } else {<a name="line.458"></a>
<span class="sourceLineNo">459</span> validateHFiles(fs, args.get(ImportTsv.BULK_OUTPUT_CONF_KEY), family, expectedKVCount);<a name="line.459"></a>
<span class="sourceLineNo">460</span> }<a name="line.460"></a>
<span class="sourceLineNo">461</span> } else {<a name="line.461"></a>
<span class="sourceLineNo">462</span> validateTable(conf, table, family, valueMultiplier, isDryRun);<a name="line.462"></a>
<span class="sourceLineNo">463</span> }<a name="line.463"></a>
<span class="sourceLineNo">464</span><a name="line.464"></a>
<span class="sourceLineNo">465</span> if (conf.getBoolean(DELETE_AFTER_LOAD_CONF, true)) {<a name="line.465"></a>
<span class="sourceLineNo">466</span> LOG.debug("Deleting test subdirectory");<a name="line.466"></a>
<span class="sourceLineNo">467</span> util.cleanupDataTestDirOnTestFS(table.getNameAsString());<a name="line.467"></a>
<span class="sourceLineNo">468</span> }<a name="line.468"></a>
<span class="sourceLineNo">469</span> return tool;<a name="line.469"></a>
<span class="sourceLineNo">470</span> }<a name="line.470"></a>
<span class="sourceLineNo">471</span><a name="line.471"></a>
<span class="sourceLineNo">472</span> /**<a name="line.472"></a>
<span class="sourceLineNo">473</span> * Confirm ImportTsv via data in online table.<a name="line.473"></a>
<span class="sourceLineNo">474</span> */<a name="line.474"></a>
<span class="sourceLineNo">475</span> private static void validateTable(Configuration conf, TableName tableName, String family,<a name="line.475"></a>
<span class="sourceLineNo">476</span> int valueMultiplier, boolean isDryRun) throws IOException {<a name="line.476"></a>
<span class="sourceLineNo">477</span><a name="line.477"></a>
<span class="sourceLineNo">478</span> LOG.debug("Validating table.");<a name="line.478"></a>
<span class="sourceLineNo">479</span> Connection connection = ConnectionFactory.createConnection(conf);<a name="line.479"></a>
<span class="sourceLineNo">480</span> Table table = connection.getTable(tableName);<a name="line.480"></a>
<span class="sourceLineNo">481</span> boolean verified = false;<a name="line.481"></a>
<span class="sourceLineNo">482</span> long pause = conf.getLong("hbase.client.pause", 5 * 1000);<a name="line.482"></a>
<span class="sourceLineNo">483</span> int numRetries = conf.getInt(HConstants.HBASE_CLIENT_RETRIES_NUMBER, 5);<a name="line.483"></a>
<span class="sourceLineNo">484</span> for (int i = 0; i &lt; numRetries; i++) {<a name="line.484"></a>
<span class="sourceLineNo">485</span> try {<a name="line.485"></a>
<span class="sourceLineNo">486</span> Scan scan = new Scan();<a name="line.486"></a>
<span class="sourceLineNo">487</span> // Scan entire family.<a name="line.487"></a>
<span class="sourceLineNo">488</span> scan.addFamily(Bytes.toBytes(family));<a name="line.488"></a>
<span class="sourceLineNo">489</span> ResultScanner resScanner = table.getScanner(scan);<a name="line.489"></a>
<span class="sourceLineNo">490</span> int numRows = 0;<a name="line.490"></a>
<span class="sourceLineNo">491</span> for (Result res : resScanner) {<a name="line.491"></a>
<span class="sourceLineNo">492</span> numRows++;<a name="line.492"></a>
<span class="sourceLineNo">493</span> assertEquals(2, res.size());<a name="line.493"></a>
<span class="sourceLineNo">494</span> List&lt;Cell&gt; kvs = res.listCells();<a name="line.494"></a>
<span class="sourceLineNo">495</span> assertTrue(CellUtil.matchingRows(kvs.get(0), Bytes.toBytes("KEY")));<a name="line.495"></a>
<span class="sourceLineNo">496</span> assertTrue(CellUtil.matchingRows(kvs.get(1), Bytes.toBytes("KEY")));<a name="line.496"></a>
<span class="sourceLineNo">497</span> assertTrue(CellUtil.matchingValue(kvs.get(0), Bytes.toBytes("VALUE" + valueMultiplier)));<a name="line.497"></a>
<span class="sourceLineNo">498</span> assertTrue(<a name="line.498"></a>
<span class="sourceLineNo">499</span> CellUtil.matchingValue(kvs.get(1), Bytes.toBytes("VALUE" + 2 * valueMultiplier)));<a name="line.499"></a>
<span class="sourceLineNo">500</span> // Only one result set is expected, so let it loop.<a name="line.500"></a>
<span class="sourceLineNo">501</span> }<a name="line.501"></a>
<span class="sourceLineNo">502</span> if (isDryRun) {<a name="line.502"></a>
<span class="sourceLineNo">503</span> assertEquals(0, numRows);<a name="line.503"></a>
<span class="sourceLineNo">504</span> } else {<a name="line.504"></a>
<span class="sourceLineNo">505</span> assertEquals(1, numRows);<a name="line.505"></a>
<span class="sourceLineNo">506</span> }<a name="line.506"></a>
<span class="sourceLineNo">507</span> verified = true;<a name="line.507"></a>
<span class="sourceLineNo">508</span> break;<a name="line.508"></a>
<span class="sourceLineNo">509</span> } catch (NullPointerException e) {<a name="line.509"></a>
<span class="sourceLineNo">510</span> // If here, a cell was empty. Presume its because updates came in<a name="line.510"></a>
<span class="sourceLineNo">511</span> // after the scanner had been opened. Wait a while and retry.<a name="line.511"></a>
<span class="sourceLineNo">512</span> }<a name="line.512"></a>
<span class="sourceLineNo">513</span> try {<a name="line.513"></a>
<span class="sourceLineNo">514</span> Thread.sleep(pause);<a name="line.514"></a>
<span class="sourceLineNo">515</span> } catch (InterruptedException e) {<a name="line.515"></a>
<span class="sourceLineNo">516</span> // continue<a name="line.516"></a>
<span class="sourceLineNo">517</span> }<a name="line.517"></a>
<span class="sourceLineNo">518</span> }<a name="line.518"></a>
<span class="sourceLineNo">519</span> table.close();<a name="line.519"></a>
<span class="sourceLineNo">520</span> connection.close();<a name="line.520"></a>
<span class="sourceLineNo">521</span> assertTrue(verified);<a name="line.521"></a>
<span class="sourceLineNo">522</span> }<a name="line.522"></a>
<span class="sourceLineNo">523</span><a name="line.523"></a>
<span class="sourceLineNo">524</span> /**<a name="line.524"></a>
<span class="sourceLineNo">525</span> * Confirm ImportTsv via HFiles on fs.<a name="line.525"></a>
<span class="sourceLineNo">526</span> */<a name="line.526"></a>
<span class="sourceLineNo">527</span> private static void validateHFiles(FileSystem fs, String outputPath, String family,<a name="line.527"></a>
<span class="sourceLineNo">528</span> int expectedKVCount) throws IOException {<a name="line.528"></a>
<span class="sourceLineNo">529</span> // validate number and content of output columns<a name="line.529"></a>
<span class="sourceLineNo">530</span> LOG.debug("Validating HFiles.");<a name="line.530"></a>
<span class="sourceLineNo">531</span> Set&lt;String&gt; configFamilies = new HashSet&lt;&gt;();<a name="line.531"></a>
<span class="sourceLineNo">532</span> configFamilies.add(family);<a name="line.532"></a>
<span class="sourceLineNo">533</span> Set&lt;String&gt; foundFamilies = new HashSet&lt;&gt;();<a name="line.533"></a>
<span class="sourceLineNo">534</span> int actualKVCount = 0;<a name="line.534"></a>
<span class="sourceLineNo">535</span> for (FileStatus cfStatus : fs.listStatus(new Path(outputPath), new OutputFilesFilter())) {<a name="line.535"></a>
<span class="sourceLineNo">536</span> String[] elements = cfStatus.getPath().toString().split(Path.SEPARATOR);<a name="line.536"></a>
<span class="sourceLineNo">537</span> String cf = elements[elements.length - 1];<a name="line.537"></a>
<span class="sourceLineNo">538</span> foundFamilies.add(cf);<a name="line.538"></a>
<span class="sourceLineNo">539</span> assertTrue(String.format(<a name="line.539"></a>
<span class="sourceLineNo">540</span> "HFile output contains a column family (%s) not present in input families (%s)", cf,<a name="line.540"></a>
<span class="sourceLineNo">541</span> configFamilies), configFamilies.contains(cf));<a name="line.541"></a>
<span class="sourceLineNo">542</span> for (FileStatus hfile : fs.listStatus(cfStatus.getPath())) {<a name="line.542"></a>
<span class="sourceLineNo">543</span> assertTrue(String.format("HFile %s appears to contain no data.", hfile.getPath()),<a name="line.543"></a>
<span class="sourceLineNo">544</span> hfile.getLen() &gt; 0);<a name="line.544"></a>
<span class="sourceLineNo">545</span> // count the number of KVs from all the hfiles<a name="line.545"></a>
<span class="sourceLineNo">546</span> if (expectedKVCount &gt; -1) {<a name="line.546"></a>
<span class="sourceLineNo">547</span> actualKVCount += getKVCountFromHfile(fs, hfile.getPath());<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> }<a name="line.550"></a>
<span class="sourceLineNo">551</span> assertTrue(String.format("HFile output does not contain the input family '%s'.", family),<a name="line.551"></a>
<span class="sourceLineNo">552</span> foundFamilies.contains(family));<a name="line.552"></a>
<span class="sourceLineNo">553</span> if (expectedKVCount &gt; -1) {<a name="line.553"></a>
<span class="sourceLineNo">554</span> assertTrue(<a name="line.554"></a>
<span class="sourceLineNo">555</span> String.format("KV count in ouput hfile=&lt;%d&gt; doesn't match with expected KV count=&lt;%d&gt;",<a name="line.555"></a>
<span class="sourceLineNo">556</span> actualKVCount, expectedKVCount),<a name="line.556"></a>
<span class="sourceLineNo">557</span> actualKVCount == expectedKVCount);<a name="line.557"></a>
<span class="sourceLineNo">558</span> }<a name="line.558"></a>
<span class="sourceLineNo">559</span> }<a name="line.559"></a>
<span class="sourceLineNo">560</span><a name="line.560"></a>
<span class="sourceLineNo">561</span> /**<a name="line.561"></a>
<span class="sourceLineNo">562</span> * Method returns the total KVs in given hfile<a name="line.562"></a>
<span class="sourceLineNo">563</span> * @param fs File System<a name="line.563"></a>
<span class="sourceLineNo">564</span> * @param p HFile path<a name="line.564"></a>
<span class="sourceLineNo">565</span> * @return KV count in the given hfile<a name="line.565"></a>
<span class="sourceLineNo">566</span> */<a name="line.566"></a>
<span class="sourceLineNo">567</span> private static int getKVCountFromHfile(FileSystem fs, Path p) throws IOException {<a name="line.567"></a>
<span class="sourceLineNo">568</span> Configuration conf = util.getConfiguration();<a name="line.568"></a>
<span class="sourceLineNo">569</span> HFile.Reader reader = HFile.createReader(fs, p, new CacheConfig(conf), true, conf);<a name="line.569"></a>
<span class="sourceLineNo">570</span> HFileScanner scanner = reader.getScanner(conf, false, false);<a name="line.570"></a>
<span class="sourceLineNo">571</span> scanner.seekTo();<a name="line.571"></a>
<span class="sourceLineNo">572</span> int count = 0;<a name="line.572"></a>
<span class="sourceLineNo">573</span> do {<a name="line.573"></a>
<span class="sourceLineNo">574</span> count++;<a name="line.574"></a>
<span class="sourceLineNo">575</span> } while (scanner.next());<a name="line.575"></a>
<span class="sourceLineNo">576</span> reader.close();<a name="line.576"></a>
<span class="sourceLineNo">577</span> return count;<a name="line.577"></a>
<span class="sourceLineNo">578</span> }<a name="line.578"></a>
<span class="sourceLineNo">579</span>}<a name="line.579"></a>
</pre>
</div>
</body>
</html>