http://git-wip-us.apache.org/repos/asf/hbase-site/blob/75eda567/apidocs/src-html/org/apache/hadoop/hbase/mapreduce/Import.html ---------------------------------------------------------------------- diff --git a/apidocs/src-html/org/apache/hadoop/hbase/mapreduce/Import.html b/apidocs/src-html/org/apache/hadoop/hbase/mapreduce/Import.html index 11b3057..fff3fe9 100644 --- a/apidocs/src-html/org/apache/hadoop/hbase/mapreduce/Import.html +++ b/apidocs/src-html/org/apache/hadoop/hbase/mapreduce/Import.html @@ -150,639 +150,643 @@ <span class="sourceLineNo">142</span> }<a name="line.142"></a> <span class="sourceLineNo">143</span><a name="line.143"></a> <span class="sourceLineNo">144</span> @Override<a name="line.144"></a> -<span class="sourceLineNo">145</span> public int compareTo(KeyValueWritableComparable o) {<a name="line.145"></a> -<span class="sourceLineNo">146</span> return CellComparator.COMPARATOR.compare(this.kv, ((KeyValueWritableComparable)o).kv);<a name="line.146"></a> -<span class="sourceLineNo">147</span> }<a name="line.147"></a> -<span class="sourceLineNo">148</span> <a name="line.148"></a> -<span class="sourceLineNo">149</span> public static class KeyValueWritableComparator extends WritableComparator {<a name="line.149"></a> -<span class="sourceLineNo">150</span><a name="line.150"></a> -<span class="sourceLineNo">151</span> @Override<a name="line.151"></a> -<span class="sourceLineNo">152</span> public int compare(byte[] b1, int s1, int l1, byte[] b2, int s2, int l2) {<a name="line.152"></a> -<span class="sourceLineNo">153</span> try {<a name="line.153"></a> -<span class="sourceLineNo">154</span> KeyValueWritableComparable kv1 = new KeyValueWritableComparable();<a name="line.154"></a> -<span class="sourceLineNo">155</span> kv1.readFields(new DataInputStream(new ByteArrayInputStream(b1, s1, l1)));<a name="line.155"></a> -<span class="sourceLineNo">156</span> KeyValueWritableComparable kv2 = new KeyValueWritableComparable();<a name="line.156"></a> -<span class="sourceLineNo">157</span> kv2.readFields(new DataInputStream(new ByteArrayInputStream(b2, s2, l2)));<a name="line.157"></a> -<span class="sourceLineNo">158</span> return compare(kv1, kv2);<a name="line.158"></a> -<span class="sourceLineNo">159</span> } catch (IOException e) {<a name="line.159"></a> -<span class="sourceLineNo">160</span> throw new RuntimeException(e);<a name="line.160"></a> -<span class="sourceLineNo">161</span> } <a name="line.161"></a> -<span class="sourceLineNo">162</span> }<a name="line.162"></a> -<span class="sourceLineNo">163</span> <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> }<a name="line.166"></a> -<span class="sourceLineNo">167</span> <a name="line.167"></a> -<span class="sourceLineNo">168</span> public static class KeyValueReducer<a name="line.168"></a> -<span class="sourceLineNo">169</span> extends<a name="line.169"></a> -<span class="sourceLineNo">170</span> Reducer<KeyValueWritableComparable, KeyValue, ImmutableBytesWritable, KeyValue> {<a name="line.170"></a> -<span class="sourceLineNo">171</span> protected void reduce(<a name="line.171"></a> -<span class="sourceLineNo">172</span> KeyValueWritableComparable row,<a name="line.172"></a> -<span class="sourceLineNo">173</span> Iterable<KeyValue> kvs,<a name="line.173"></a> -<span class="sourceLineNo">174</span> Reducer<KeyValueWritableComparable,<a name="line.174"></a> -<span class="sourceLineNo">175</span> KeyValue, ImmutableBytesWritable, KeyValue>.Context context)<a name="line.175"></a> -<span class="sourceLineNo">176</span> throws java.io.IOException, InterruptedException {<a name="line.176"></a> -<span class="sourceLineNo">177</span> int index = 0;<a name="line.177"></a> -<span class="sourceLineNo">178</span> for (KeyValue kv : kvs) {<a name="line.178"></a> -<span class="sourceLineNo">179</span> context.write(new ImmutableBytesWritable(kv.getRowArray()), kv);<a name="line.179"></a> -<span class="sourceLineNo">180</span> if (++index % 100 == 0)<a name="line.180"></a> -<span class="sourceLineNo">181</span> context.setStatus("Wrote " + index + " KeyValues, "<a name="line.181"></a> -<span class="sourceLineNo">182</span> + "and the rowkey whose is being wrote is " + Bytes.toString(kv.getRowArray())); <a name="line.182"></a> -<span class="sourceLineNo">183</span> }<a name="line.183"></a> -<span class="sourceLineNo">184</span> }<a name="line.184"></a> -<span class="sourceLineNo">185</span> }<a name="line.185"></a> -<span class="sourceLineNo">186</span> <a name="line.186"></a> -<span class="sourceLineNo">187</span> public static class KeyValueSortImporter <a name="line.187"></a> -<span class="sourceLineNo">188</span> extends TableMapper<KeyValueWritableComparable, KeyValue> {<a name="line.188"></a> -<span class="sourceLineNo">189</span> private Map<byte[], byte[]> cfRenameMap;<a name="line.189"></a> -<span class="sourceLineNo">190</span> private Filter filter;<a name="line.190"></a> -<span class="sourceLineNo">191</span> private static final Log LOG = LogFactory.getLog(KeyValueImporter.class);<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> * @param row The current table row key.<a name="line.194"></a> -<span class="sourceLineNo">195</span> * @param value The columns.<a name="line.195"></a> -<span class="sourceLineNo">196</span> * @param context The current context.<a name="line.196"></a> -<span class="sourceLineNo">197</span> * @throws IOException When something is broken with the data.<a name="line.197"></a> -<span class="sourceLineNo">198</span> */<a name="line.198"></a> -<span class="sourceLineNo">199</span> @Override<a name="line.199"></a> -<span class="sourceLineNo">200</span> public void map(ImmutableBytesWritable row, Result value,<a name="line.200"></a> -<span class="sourceLineNo">201</span> Context context)<a name="line.201"></a> -<span class="sourceLineNo">202</span> throws IOException {<a name="line.202"></a> -<span class="sourceLineNo">203</span> try {<a name="line.203"></a> -<span class="sourceLineNo">204</span> if (LOG.isTraceEnabled()) {<a name="line.204"></a> -<span class="sourceLineNo">205</span> LOG.trace("Considering the row."<a name="line.205"></a> -<span class="sourceLineNo">206</span> + Bytes.toString(row.get(), row.getOffset(), row.getLength()));<a name="line.206"></a> -<span class="sourceLineNo">207</span> }<a name="line.207"></a> -<span class="sourceLineNo">208</span> if (filter == null<a name="line.208"></a> -<span class="sourceLineNo">209</span> || !filter.filterRowKey(KeyValueUtil.createFirstOnRow(row.get(), row.getOffset(),<a name="line.209"></a> -<span class="sourceLineNo">210</span> (short) row.getLength()))) {<a name="line.210"></a> -<span class="sourceLineNo">211</span> for (Cell kv : value.rawCells()) {<a name="line.211"></a> -<span class="sourceLineNo">212</span> kv = filterKv(filter, kv);<a name="line.212"></a> -<span class="sourceLineNo">213</span> // skip if we filtered it out<a name="line.213"></a> -<span class="sourceLineNo">214</span> if (kv == null) continue;<a name="line.214"></a> -<span class="sourceLineNo">215</span> // TODO get rid of ensureKeyValue<a name="line.215"></a> -<span class="sourceLineNo">216</span> KeyValue ret = KeyValueUtil.ensureKeyValue(convertKv(kv, cfRenameMap));<a name="line.216"></a> -<span class="sourceLineNo">217</span> context.write(new KeyValueWritableComparable(ret.createKeyOnly(false)), ret); <a name="line.217"></a> -<span class="sourceLineNo">218</span> }<a name="line.218"></a> -<span class="sourceLineNo">219</span> }<a name="line.219"></a> -<span class="sourceLineNo">220</span> } catch (InterruptedException e) {<a name="line.220"></a> -<span class="sourceLineNo">221</span> e.printStackTrace();<a name="line.221"></a> -<span class="sourceLineNo">222</span> }<a name="line.222"></a> -<span class="sourceLineNo">223</span> }<a name="line.223"></a> -<span class="sourceLineNo">224</span> <a name="line.224"></a> -<span class="sourceLineNo">225</span> @Override<a name="line.225"></a> -<span class="sourceLineNo">226</span> public void setup(Context context) throws IOException { <a name="line.226"></a> -<span class="sourceLineNo">227</span> cfRenameMap = createCfRenameMap(context.getConfiguration());<a name="line.227"></a> -<span class="sourceLineNo">228</span> filter = instantiateFilter(context.getConfiguration());<a name="line.228"></a> -<span class="sourceLineNo">229</span> int reduceNum = context.getNumReduceTasks();<a name="line.229"></a> -<span class="sourceLineNo">230</span> Configuration conf = context.getConfiguration();<a name="line.230"></a> -<span class="sourceLineNo">231</span> TableName tableName = TableName.valueOf(context.getConfiguration().get(TABLE_NAME));<a name="line.231"></a> -<span class="sourceLineNo">232</span> try (Connection conn = ConnectionFactory.createConnection(conf); <a name="line.232"></a> -<span class="sourceLineNo">233</span> RegionLocator regionLocator = conn.getRegionLocator(tableName)) {<a name="line.233"></a> -<span class="sourceLineNo">234</span> byte[][] startKeys = regionLocator.getStartKeys();<a name="line.234"></a> -<span class="sourceLineNo">235</span> if (startKeys.length != reduceNum) {<a name="line.235"></a> -<span class="sourceLineNo">236</span> throw new IOException("Region split after job initialization");<a name="line.236"></a> -<span class="sourceLineNo">237</span> }<a name="line.237"></a> -<span class="sourceLineNo">238</span> KeyValueWritableComparable[] startKeyWraps = <a name="line.238"></a> -<span class="sourceLineNo">239</span> new KeyValueWritableComparable[startKeys.length - 1];<a name="line.239"></a> -<span class="sourceLineNo">240</span> for (int i = 1; i < startKeys.length; ++i) {<a name="line.240"></a> -<span class="sourceLineNo">241</span> startKeyWraps[i - 1] = <a name="line.241"></a> -<span class="sourceLineNo">242</span> new KeyValueWritableComparable(KeyValueUtil.createFirstOnRow(startKeys[i]));<a name="line.242"></a> -<span class="sourceLineNo">243</span> }<a name="line.243"></a> -<span class="sourceLineNo">244</span> KeyValueWritableComparablePartitioner.START_KEYS = startKeyWraps;<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> <a name="line.248"></a> -<span class="sourceLineNo">249</span> /**<a name="line.249"></a> -<span class="sourceLineNo">250</span> * A mapper that just writes out KeyValues.<a name="line.250"></a> -<span class="sourceLineNo">251</span> */<a name="line.251"></a> -<span class="sourceLineNo">252</span> public static class KeyValueImporter extends TableMapper<ImmutableBytesWritable, KeyValue> {<a name="line.252"></a> -<span class="sourceLineNo">253</span> private Map<byte[], byte[]> cfRenameMap;<a name="line.253"></a> -<span class="sourceLineNo">254</span> private Filter filter;<a name="line.254"></a> -<span class="sourceLineNo">255</span> private static final Log LOG = LogFactory.getLog(KeyValueImporter.class);<a name="line.255"></a> -<span class="sourceLineNo">256</span><a name="line.256"></a> -<span class="sourceLineNo">257</span> /**<a name="line.257"></a> -<span class="sourceLineNo">258</span> * @param row The current table row key.<a name="line.258"></a> -<span class="sourceLineNo">259</span> * @param value The columns.<a name="line.259"></a> -<span class="sourceLineNo">260</span> * @param context The current context.<a name="line.260"></a> -<span class="sourceLineNo">261</span> * @throws IOException When something is broken with the data.<a name="line.261"></a> -<span class="sourceLineNo">262</span> */<a name="line.262"></a> -<span class="sourceLineNo">263</span> @Override<a name="line.263"></a> -<span class="sourceLineNo">264</span> public void map(ImmutableBytesWritable row, Result value,<a name="line.264"></a> -<span class="sourceLineNo">265</span> Context context)<a name="line.265"></a> -<span class="sourceLineNo">266</span> throws IOException {<a name="line.266"></a> -<span class="sourceLineNo">267</span> try {<a name="line.267"></a> -<span class="sourceLineNo">268</span> if (LOG.isTraceEnabled()) {<a name="line.268"></a> -<span class="sourceLineNo">269</span> LOG.trace("Considering the row."<a name="line.269"></a> -<span class="sourceLineNo">270</span> + Bytes.toString(row.get(), row.getOffset(), row.getLength()));<a name="line.270"></a> -<span class="sourceLineNo">271</span> }<a name="line.271"></a> -<span class="sourceLineNo">272</span> if (filter == null<a name="line.272"></a> -<span class="sourceLineNo">273</span> || !filter.filterRowKey(KeyValueUtil.createFirstOnRow(row.get(), row.getOffset(),<a name="line.273"></a> -<span class="sourceLineNo">274</span> (short) row.getLength()))) {<a name="line.274"></a> -<span class="sourceLineNo">275</span> for (Cell kv : value.rawCells()) {<a name="line.275"></a> -<span class="sourceLineNo">276</span> kv = filterKv(filter, kv);<a name="line.276"></a> -<span class="sourceLineNo">277</span> // skip if we filtered it out<a name="line.277"></a> -<span class="sourceLineNo">278</span> if (kv == null) continue;<a name="line.278"></a> -<span class="sourceLineNo">279</span> // TODO get rid of ensureKeyValue<a name="line.279"></a> -<span class="sourceLineNo">280</span> context.write(row, KeyValueUtil.ensureKeyValue(convertKv(kv, cfRenameMap)));<a name="line.280"></a> -<span class="sourceLineNo">281</span> }<a name="line.281"></a> -<span class="sourceLineNo">282</span> }<a name="line.282"></a> -<span class="sourceLineNo">283</span> } catch (InterruptedException e) {<a name="line.283"></a> -<span class="sourceLineNo">284</span> e.printStackTrace();<a name="line.284"></a> -<span class="sourceLineNo">285</span> }<a name="line.285"></a> -<span class="sourceLineNo">286</span> }<a name="line.286"></a> -<span class="sourceLineNo">287</span><a name="line.287"></a> -<span class="sourceLineNo">288</span> @Override<a name="line.288"></a> -<span class="sourceLineNo">289</span> public void setup(Context context) {<a name="line.289"></a> -<span class="sourceLineNo">290</span> cfRenameMap = createCfRenameMap(context.getConfiguration());<a name="line.290"></a> -<span class="sourceLineNo">291</span> filter = instantiateFilter(context.getConfiguration());<a name="line.291"></a> -<span class="sourceLineNo">292</span> }<a name="line.292"></a> -<span class="sourceLineNo">293</span> }<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> * Write table content out to files in hdfs.<a name="line.296"></a> -<span class="sourceLineNo">297</span> */<a name="line.297"></a> -<span class="sourceLineNo">298</span> public static class Importer extends TableMapper<ImmutableBytesWritable, Mutation> {<a name="line.298"></a> -<span class="sourceLineNo">299</span> private Map<byte[], byte[]> cfRenameMap;<a name="line.299"></a> -<span class="sourceLineNo">300</span> private List<UUID> clusterIds;<a name="line.300"></a> -<span class="sourceLineNo">301</span> private Filter filter;<a name="line.301"></a> -<span class="sourceLineNo">302</span> private Durability durability;<a name="line.302"></a> -<span class="sourceLineNo">303</span><a name="line.303"></a> -<span class="sourceLineNo">304</span> /**<a name="line.304"></a> -<span class="sourceLineNo">305</span> * @param row The current table row key.<a name="line.305"></a> -<span class="sourceLineNo">306</span> * @param value The columns.<a name="line.306"></a> -<span class="sourceLineNo">307</span> * @param context The current context.<a name="line.307"></a> -<span class="sourceLineNo">308</span> * @throws IOException When something is broken with the data.<a name="line.308"></a> -<span class="sourceLineNo">309</span> */<a name="line.309"></a> -<span class="sourceLineNo">310</span> @Override<a name="line.310"></a> -<span class="sourceLineNo">311</span> public void map(ImmutableBytesWritable row, Result value,<a name="line.311"></a> -<span class="sourceLineNo">312</span> Context context)<a name="line.312"></a> -<span class="sourceLineNo">313</span> throws IOException {<a name="line.313"></a> -<span class="sourceLineNo">314</span> try {<a name="line.314"></a> -<span class="sourceLineNo">315</span> writeResult(row, value, context);<a name="line.315"></a> -<span class="sourceLineNo">316</span> } catch (InterruptedException e) {<a name="line.316"></a> -<span class="sourceLineNo">317</span> e.printStackTrace();<a name="line.317"></a> -<span class="sourceLineNo">318</span> }<a name="line.318"></a> -<span class="sourceLineNo">319</span> }<a name="line.319"></a> -<span class="sourceLineNo">320</span><a name="line.320"></a> -<span class="sourceLineNo">321</span> private void writeResult(ImmutableBytesWritable key, Result result, Context context)<a name="line.321"></a> -<span class="sourceLineNo">322</span> throws IOException, InterruptedException {<a name="line.322"></a> -<span class="sourceLineNo">323</span> Put put = null;<a name="line.323"></a> -<span class="sourceLineNo">324</span> Delete delete = null;<a name="line.324"></a> -<span class="sourceLineNo">325</span> if (LOG.isTraceEnabled()) {<a name="line.325"></a> -<span class="sourceLineNo">326</span> LOG.trace("Considering the row."<a name="line.326"></a> -<span class="sourceLineNo">327</span> + Bytes.toString(key.get(), key.getOffset(), key.getLength()));<a name="line.327"></a> -<span class="sourceLineNo">328</span> }<a name="line.328"></a> -<span class="sourceLineNo">329</span> if (filter == null<a name="line.329"></a> -<span class="sourceLineNo">330</span> || !filter.filterRowKey(KeyValueUtil.createFirstOnRow(key.get(), key.getOffset(),<a name="line.330"></a> -<span class="sourceLineNo">331</span> (short) key.getLength()))) {<a name="line.331"></a> -<span class="sourceLineNo">332</span> processKV(key, result, context, put, delete);<a name="line.332"></a> -<span class="sourceLineNo">333</span> }<a name="line.333"></a> -<span class="sourceLineNo">334</span> }<a name="line.334"></a> -<span class="sourceLineNo">335</span><a name="line.335"></a> -<span class="sourceLineNo">336</span> protected void processKV(ImmutableBytesWritable key, Result result, Context context, Put put,<a name="line.336"></a> -<span class="sourceLineNo">337</span> Delete delete) throws IOException, InterruptedException {<a name="line.337"></a> -<span class="sourceLineNo">338</span> for (Cell kv : result.rawCells()) {<a name="line.338"></a> -<span class="sourceLineNo">339</span> kv = filterKv(filter, kv);<a name="line.339"></a> -<span class="sourceLineNo">340</span> // skip if we filter it out<a name="line.340"></a> -<span class="sourceLineNo">341</span> if (kv == null) continue;<a name="line.341"></a> -<span class="sourceLineNo">342</span><a name="line.342"></a> -<span class="sourceLineNo">343</span> kv = convertKv(kv, cfRenameMap);<a name="line.343"></a> -<span class="sourceLineNo">344</span> // Deletes and Puts are gathered and written when finished<a name="line.344"></a> -<span class="sourceLineNo">345</span> /*<a name="line.345"></a> -<span class="sourceLineNo">346</span> * If there are sequence of mutations and tombstones in an Export, and after Import the same<a name="line.346"></a> -<span class="sourceLineNo">347</span> * sequence should be restored as it is. If we combine all Delete tombstones into single<a name="line.347"></a> -<span class="sourceLineNo">348</span> * request then there is chance of ignoring few DeleteFamily tombstones, because if we<a name="line.348"></a> -<span class="sourceLineNo">349</span> * submit multiple DeleteFamily tombstones in single Delete request then we are maintaining<a name="line.349"></a> -<span class="sourceLineNo">350</span> * only newest in hbase table and ignoring other. Check - HBASE-12065<a name="line.350"></a> -<span class="sourceLineNo">351</span> */<a name="line.351"></a> -<span class="sourceLineNo">352</span> if (CellUtil.isDeleteFamily(kv)) {<a name="line.352"></a> -<span class="sourceLineNo">353</span> Delete deleteFamily = new Delete(key.get());<a name="line.353"></a> -<span class="sourceLineNo">354</span> deleteFamily.addDeleteMarker(kv);<a name="line.354"></a> -<span class="sourceLineNo">355</span> if (durability != null) {<a name="line.355"></a> -<span class="sourceLineNo">356</span> deleteFamily.setDurability(durability);<a name="line.356"></a> -<span class="sourceLineNo">357</span> }<a name="line.357"></a> -<span class="sourceLineNo">358</span> deleteFamily.setClusterIds(clusterIds);<a name="line.358"></a> -<span class="sourceLineNo">359</span> context.write(key, deleteFamily);<a name="line.359"></a> -<span class="sourceLineNo">360</span> } else if (CellUtil.isDelete(kv)) {<a name="line.360"></a> -<span class="sourceLineNo">361</span> if (delete == null) {<a name="line.361"></a> -<span class="sourceLineNo">362</span> delete = new Delete(key.get());<a name="line.362"></a> -<span class="sourceLineNo">363</span> }<a name="line.363"></a> -<span class="sourceLineNo">364</span> delete.addDeleteMarker(kv);<a name="line.364"></a> -<span class="sourceLineNo">365</span> } else {<a name="line.365"></a> -<span class="sourceLineNo">366</span> if (put == null) {<a name="line.366"></a> -<span class="sourceLineNo">367</span> put = new Put(key.get());<a name="line.367"></a> -<span class="sourceLineNo">368</span> }<a name="line.368"></a> -<span class="sourceLineNo">369</span> addPutToKv(put, kv);<a name="line.369"></a> -<span class="sourceLineNo">370</span> }<a name="line.370"></a> -<span class="sourceLineNo">371</span> }<a name="line.371"></a> -<span class="sourceLineNo">372</span> if (put != null) {<a name="line.372"></a> -<span class="sourceLineNo">373</span> if (durability != null) {<a name="line.373"></a> -<span class="sourceLineNo">374</span> put.setDurability(durability);<a name="line.374"></a> -<span class="sourceLineNo">375</span> }<a name="line.375"></a> -<span class="sourceLineNo">376</span> put.setClusterIds(clusterIds);<a name="line.376"></a> -<span class="sourceLineNo">377</span> context.write(key, put);<a name="line.377"></a> -<span class="sourceLineNo">378</span> }<a name="line.378"></a> -<span class="sourceLineNo">379</span> if (delete != null) {<a name="line.379"></a> -<span class="sourceLineNo">380</span> if (durability != null) {<a name="line.380"></a> -<span class="sourceLineNo">381</span> delete.setDurability(durability);<a name="line.381"></a> -<span class="sourceLineNo">382</span> }<a name="line.382"></a> -<span class="sourceLineNo">383</span> delete.setClusterIds(clusterIds);<a name="line.383"></a> -<span class="sourceLineNo">384</span> context.write(key, delete);<a name="line.384"></a> -<span class="sourceLineNo">385</span> }<a name="line.385"></a> -<span class="sourceLineNo">386</span> }<a name="line.386"></a> -<span class="sourceLineNo">387</span><a name="line.387"></a> -<span class="sourceLineNo">388</span> protected void addPutToKv(Put put, Cell kv) throws IOException {<a name="line.388"></a> -<span class="sourceLineNo">389</span> put.add(kv);<a name="line.389"></a> +<span class="sourceLineNo">145</span> @edu.umd.cs.findbugs.annotations.SuppressWarnings(value="EQ_COMPARETO_USE_OBJECT_EQUALS",<a name="line.145"></a> +<span class="sourceLineNo">146</span> justification="This is wrong, yes, but we should be purging Writables, not fixing them")<a name="line.146"></a> +<span class="sourceLineNo">147</span> public int compareTo(KeyValueWritableComparable o) {<a name="line.147"></a> +<span class="sourceLineNo">148</span> return CellComparator.COMPARATOR.compare(this.kv, ((KeyValueWritableComparable)o).kv);<a name="line.148"></a> +<span class="sourceLineNo">149</span> }<a name="line.149"></a> +<span class="sourceLineNo">150</span> <a name="line.150"></a> +<span class="sourceLineNo">151</span> public static class KeyValueWritableComparator extends WritableComparator {<a name="line.151"></a> +<span class="sourceLineNo">152</span><a name="line.152"></a> +<span class="sourceLineNo">153</span> @Override<a name="line.153"></a> +<span class="sourceLineNo">154</span> public int compare(byte[] b1, int s1, int l1, byte[] b2, int s2, int l2) {<a name="line.154"></a> +<span class="sourceLineNo">155</span> try {<a name="line.155"></a> +<span class="sourceLineNo">156</span> KeyValueWritableComparable kv1 = new KeyValueWritableComparable();<a name="line.156"></a> +<span class="sourceLineNo">157</span> kv1.readFields(new DataInputStream(new ByteArrayInputStream(b1, s1, l1)));<a name="line.157"></a> +<span class="sourceLineNo">158</span> KeyValueWritableComparable kv2 = new KeyValueWritableComparable();<a name="line.158"></a> +<span class="sourceLineNo">159</span> kv2.readFields(new DataInputStream(new ByteArrayInputStream(b2, s2, l2)));<a name="line.159"></a> +<span class="sourceLineNo">160</span> return compare(kv1, kv2);<a name="line.160"></a> +<span class="sourceLineNo">161</span> } catch (IOException e) {<a name="line.161"></a> +<span class="sourceLineNo">162</span> throw new RuntimeException(e);<a name="line.162"></a> +<span class="sourceLineNo">163</span> } <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> }<a name="line.166"></a> +<span class="sourceLineNo">167</span> <a name="line.167"></a> +<span class="sourceLineNo">168</span> }<a name="line.168"></a> +<span class="sourceLineNo">169</span> <a name="line.169"></a> +<span class="sourceLineNo">170</span> public static class KeyValueReducer<a name="line.170"></a> +<span class="sourceLineNo">171</span> extends<a name="line.171"></a> +<span class="sourceLineNo">172</span> Reducer<KeyValueWritableComparable, KeyValue, ImmutableBytesWritable, KeyValue> {<a name="line.172"></a> +<span class="sourceLineNo">173</span> protected void reduce(<a name="line.173"></a> +<span class="sourceLineNo">174</span> KeyValueWritableComparable row,<a name="line.174"></a> +<span class="sourceLineNo">175</span> Iterable<KeyValue> kvs,<a name="line.175"></a> +<span class="sourceLineNo">176</span> Reducer<KeyValueWritableComparable,<a name="line.176"></a> +<span class="sourceLineNo">177</span> KeyValue, ImmutableBytesWritable, KeyValue>.Context context)<a name="line.177"></a> +<span class="sourceLineNo">178</span> throws java.io.IOException, InterruptedException {<a name="line.178"></a> +<span class="sourceLineNo">179</span> int index = 0;<a name="line.179"></a> +<span class="sourceLineNo">180</span> for (KeyValue kv : kvs) {<a name="line.180"></a> +<span class="sourceLineNo">181</span> context.write(new ImmutableBytesWritable(kv.getRowArray()), kv);<a name="line.181"></a> +<span class="sourceLineNo">182</span> if (++index % 100 == 0)<a name="line.182"></a> +<span class="sourceLineNo">183</span> context.setStatus("Wrote " + index + " KeyValues, "<a name="line.183"></a> +<span class="sourceLineNo">184</span> + "and the rowkey whose is being wrote is " + Bytes.toString(kv.getRowArray())); <a name="line.184"></a> +<span class="sourceLineNo">185</span> }<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> <a name="line.188"></a> +<span class="sourceLineNo">189</span> public static class KeyValueSortImporter <a name="line.189"></a> +<span class="sourceLineNo">190</span> extends TableMapper<KeyValueWritableComparable, KeyValue> {<a name="line.190"></a> +<span class="sourceLineNo">191</span> private Map<byte[], byte[]> cfRenameMap;<a name="line.191"></a> +<span class="sourceLineNo">192</span> private Filter filter;<a name="line.192"></a> +<span class="sourceLineNo">193</span> private static final Log LOG = LogFactory.getLog(KeyValueImporter.class);<a name="line.193"></a> +<span class="sourceLineNo">194</span><a name="line.194"></a> +<span class="sourceLineNo">195</span> /**<a name="line.195"></a> +<span class="sourceLineNo">196</span> * @param row The current table row key.<a name="line.196"></a> +<span class="sourceLineNo">197</span> * @param value The columns.<a name="line.197"></a> +<span class="sourceLineNo">198</span> * @param context The current context.<a name="line.198"></a> +<span class="sourceLineNo">199</span> * @throws IOException When something is broken with the data.<a name="line.199"></a> +<span class="sourceLineNo">200</span> */<a name="line.200"></a> +<span class="sourceLineNo">201</span> @Override<a name="line.201"></a> +<span class="sourceLineNo">202</span> public void map(ImmutableBytesWritable row, Result value,<a name="line.202"></a> +<span class="sourceLineNo">203</span> Context context)<a name="line.203"></a> +<span class="sourceLineNo">204</span> throws IOException {<a name="line.204"></a> +<span class="sourceLineNo">205</span> try {<a name="line.205"></a> +<span class="sourceLineNo">206</span> if (LOG.isTraceEnabled()) {<a name="line.206"></a> +<span class="sourceLineNo">207</span> LOG.trace("Considering the row."<a name="line.207"></a> +<span class="sourceLineNo">208</span> + Bytes.toString(row.get(), row.getOffset(), row.getLength()));<a name="line.208"></a> +<span class="sourceLineNo">209</span> }<a name="line.209"></a> +<span class="sourceLineNo">210</span> if (filter == null<a name="line.210"></a> +<span class="sourceLineNo">211</span> || !filter.filterRowKey(KeyValueUtil.createFirstOnRow(row.get(), row.getOffset(),<a name="line.211"></a> +<span class="sourceLineNo">212</span> (short) row.getLength()))) {<a name="line.212"></a> +<span class="sourceLineNo">213</span> for (Cell kv : value.rawCells()) {<a name="line.213"></a> +<span class="sourceLineNo">214</span> kv = filterKv(filter, kv);<a name="line.214"></a> +<span class="sourceLineNo">215</span> // skip if we filtered it out<a name="line.215"></a> +<span class="sourceLineNo">216</span> if (kv == null) continue;<a name="line.216"></a> +<span class="sourceLineNo">217</span> // TODO get rid of ensureKeyValue<a name="line.217"></a> +<span class="sourceLineNo">218</span> KeyValue ret = KeyValueUtil.ensureKeyValue(convertKv(kv, cfRenameMap));<a name="line.218"></a> +<span class="sourceLineNo">219</span> context.write(new KeyValueWritableComparable(ret.createKeyOnly(false)), ret); <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> } catch (InterruptedException e) {<a name="line.222"></a> +<span class="sourceLineNo">223</span> e.printStackTrace();<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> <a name="line.226"></a> +<span class="sourceLineNo">227</span> @Override<a name="line.227"></a> +<span class="sourceLineNo">228</span> public void setup(Context context) throws IOException { <a name="line.228"></a> +<span class="sourceLineNo">229</span> cfRenameMap = createCfRenameMap(context.getConfiguration());<a name="line.229"></a> +<span class="sourceLineNo">230</span> filter = instantiateFilter(context.getConfiguration());<a name="line.230"></a> +<span class="sourceLineNo">231</span> int reduceNum = context.getNumReduceTasks();<a name="line.231"></a> +<span class="sourceLineNo">232</span> Configuration conf = context.getConfiguration();<a name="line.232"></a> +<span class="sourceLineNo">233</span> TableName tableName = TableName.valueOf(context.getConfiguration().get(TABLE_NAME));<a name="line.233"></a> +<span class="sourceLineNo">234</span> try (Connection conn = ConnectionFactory.createConnection(conf); <a name="line.234"></a> +<span class="sourceLineNo">235</span> RegionLocator regionLocator = conn.getRegionLocator(tableName)) {<a name="line.235"></a> +<span class="sourceLineNo">236</span> byte[][] startKeys = regionLocator.getStartKeys();<a name="line.236"></a> +<span class="sourceLineNo">237</span> if (startKeys.length != reduceNum) {<a name="line.237"></a> +<span class="sourceLineNo">238</span> throw new IOException("Region split after job initialization");<a name="line.238"></a> +<span class="sourceLineNo">239</span> }<a name="line.239"></a> +<span class="sourceLineNo">240</span> KeyValueWritableComparable[] startKeyWraps = <a name="line.240"></a> +<span class="sourceLineNo">241</span> new KeyValueWritableComparable[startKeys.length - 1];<a name="line.241"></a> +<span class="sourceLineNo">242</span> for (int i = 1; i < startKeys.length; ++i) {<a name="line.242"></a> +<span class="sourceLineNo">243</span> startKeyWraps[i - 1] = <a name="line.243"></a> +<span class="sourceLineNo">244</span> new KeyValueWritableComparable(KeyValueUtil.createFirstOnRow(startKeys[i]));<a name="line.244"></a> +<span class="sourceLineNo">245</span> }<a name="line.245"></a> +<span class="sourceLineNo">246</span> KeyValueWritableComparablePartitioner.START_KEYS = startKeyWraps;<a name="line.246"></a> +<span class="sourceLineNo">247</span> }<a name="line.247"></a> +<span class="sourceLineNo">248</span> }<a name="line.248"></a> +<span class="sourceLineNo">249</span> }<a name="line.249"></a> +<span class="sourceLineNo">250</span> <a name="line.250"></a> +<span class="sourceLineNo">251</span> /**<a name="line.251"></a> +<span class="sourceLineNo">252</span> * A mapper that just writes out KeyValues.<a name="line.252"></a> +<span class="sourceLineNo">253</span> */<a name="line.253"></a> +<span class="sourceLineNo">254</span> @edu.umd.cs.findbugs.annotations.SuppressWarnings(value="EQ_COMPARETO_USE_OBJECT_EQUALS",<a name="line.254"></a> +<span class="sourceLineNo">255</span> justification="Writables are going away and this has been this way forever")<a name="line.255"></a> +<span class="sourceLineNo">256</span> public static class KeyValueImporter extends TableMapper<ImmutableBytesWritable, KeyValue> {<a name="line.256"></a> +<span class="sourceLineNo">257</span> private Map<byte[], byte[]> cfRenameMap;<a name="line.257"></a> +<span class="sourceLineNo">258</span> private Filter filter;<a name="line.258"></a> +<span class="sourceLineNo">259</span> private static final Log LOG = LogFactory.getLog(KeyValueImporter.class);<a name="line.259"></a> +<span class="sourceLineNo">260</span><a name="line.260"></a> +<span class="sourceLineNo">261</span> /**<a name="line.261"></a> +<span class="sourceLineNo">262</span> * @param row The current table row key.<a name="line.262"></a> +<span class="sourceLineNo">263</span> * @param value The columns.<a name="line.263"></a> +<span class="sourceLineNo">264</span> * @param context The current context.<a name="line.264"></a> +<span class="sourceLineNo">265</span> * @throws IOException When something is broken with the data.<a name="line.265"></a> +<span class="sourceLineNo">266</span> */<a name="line.266"></a> +<span class="sourceLineNo">267</span> @Override<a name="line.267"></a> +<span class="sourceLineNo">268</span> public void map(ImmutableBytesWritable row, Result value,<a name="line.268"></a> +<span class="sourceLineNo">269</span> Context context)<a name="line.269"></a> +<span class="sourceLineNo">270</span> throws IOException {<a name="line.270"></a> +<span class="sourceLineNo">271</span> try {<a name="line.271"></a> +<span class="sourceLineNo">272</span> if (LOG.isTraceEnabled()) {<a name="line.272"></a> +<span class="sourceLineNo">273</span> LOG.trace("Considering the row."<a name="line.273"></a> +<span class="sourceLineNo">274</span> + Bytes.toString(row.get(), row.getOffset(), row.getLength()));<a name="line.274"></a> +<span class="sourceLineNo">275</span> }<a name="line.275"></a> +<span class="sourceLineNo">276</span> if (filter == null<a name="line.276"></a> +<span class="sourceLineNo">277</span> || !filter.filterRowKey(KeyValueUtil.createFirstOnRow(row.get(), row.getOffset(),<a name="line.277"></a> +<span class="sourceLineNo">278</span> (short) row.getLength()))) {<a name="line.278"></a> +<span class="sourceLineNo">279</span> for (Cell kv : value.rawCells()) {<a name="line.279"></a> +<span class="sourceLineNo">280</span> kv = filterKv(filter, kv);<a name="line.280"></a> +<span class="sourceLineNo">281</span> // skip if we filtered it out<a name="line.281"></a> +<span class="sourceLineNo">282</span> if (kv == null) continue;<a name="line.282"></a> +<span class="sourceLineNo">283</span> // TODO get rid of ensureKeyValue<a name="line.283"></a> +<span class="sourceLineNo">284</span> context.write(row, KeyValueUtil.ensureKeyValue(convertKv(kv, cfRenameMap)));<a name="line.284"></a> +<span class="sourceLineNo">285</span> }<a name="line.285"></a> +<span class="sourceLineNo">286</span> }<a name="line.286"></a> +<span class="sourceLineNo">287</span> } catch (InterruptedException e) {<a name="line.287"></a> +<span class="sourceLineNo">288</span> e.printStackTrace();<a name="line.288"></a> +<span class="sourceLineNo">289</span> }<a name="line.289"></a> +<span class="sourceLineNo">290</span> }<a name="line.290"></a> +<span class="sourceLineNo">291</span><a name="line.291"></a> +<span class="sourceLineNo">292</span> @Override<a name="line.292"></a> +<span class="sourceLineNo">293</span> public void setup(Context context) {<a name="line.293"></a> +<span class="sourceLineNo">294</span> cfRenameMap = createCfRenameMap(context.getConfiguration());<a name="line.294"></a> +<span class="sourceLineNo">295</span> filter = instantiateFilter(context.getConfiguration());<a name="line.295"></a> +<span class="sourceLineNo">296</span> }<a name="line.296"></a> +<span class="sourceLineNo">297</span> }<a name="line.297"></a> +<span class="sourceLineNo">298</span><a name="line.298"></a> +<span class="sourceLineNo">299</span> /**<a name="line.299"></a> +<span class="sourceLineNo">300</span> * Write table content out to files in hdfs.<a name="line.300"></a> +<span class="sourceLineNo">301</span> */<a name="line.301"></a> +<span class="sourceLineNo">302</span> public static class Importer extends TableMapper<ImmutableBytesWritable, Mutation> {<a name="line.302"></a> +<span class="sourceLineNo">303</span> private Map<byte[], byte[]> cfRenameMap;<a name="line.303"></a> +<span class="sourceLineNo">304</span> private List<UUID> clusterIds;<a name="line.304"></a> +<span class="sourceLineNo">305</span> private Filter filter;<a name="line.305"></a> +<span class="sourceLineNo">306</span> private Durability durability;<a name="line.306"></a> +<span class="sourceLineNo">307</span><a name="line.307"></a> +<span class="sourceLineNo">308</span> /**<a name="line.308"></a> +<span class="sourceLineNo">309</span> * @param row The current table row key.<a name="line.309"></a> +<span class="sourceLineNo">310</span> * @param value The columns.<a name="line.310"></a> +<span class="sourceLineNo">311</span> * @param context The current context.<a name="line.311"></a> +<span class="sourceLineNo">312</span> * @throws IOException When something is broken with the data.<a name="line.312"></a> +<span class="sourceLineNo">313</span> */<a name="line.313"></a> +<span class="sourceLineNo">314</span> @Override<a name="line.314"></a> +<span class="sourceLineNo">315</span> public void map(ImmutableBytesWritable row, Result value,<a name="line.315"></a> +<span class="sourceLineNo">316</span> Context context)<a name="line.316"></a> +<span class="sourceLineNo">317</span> throws IOException {<a name="line.317"></a> +<span class="sourceLineNo">318</span> try {<a name="line.318"></a> +<span class="sourceLineNo">319</span> writeResult(row, value, context);<a name="line.319"></a> +<span class="sourceLineNo">320</span> } catch (InterruptedException e) {<a name="line.320"></a> +<span class="sourceLineNo">321</span> e.printStackTrace();<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><a name="line.324"></a> +<span class="sourceLineNo">325</span> private void writeResult(ImmutableBytesWritable key, Result result, Context context)<a name="line.325"></a> +<span class="sourceLineNo">326</span> throws IOException, InterruptedException {<a name="line.326"></a> +<span class="sourceLineNo">327</span> Put put = null;<a name="line.327"></a> +<span class="sourceLineNo">328</span> Delete delete = null;<a name="line.328"></a> +<span class="sourceLineNo">329</span> if (LOG.isTraceEnabled()) {<a name="line.329"></a> +<span class="sourceLineNo">330</span> LOG.trace("Considering the row."<a name="line.330"></a> +<span class="sourceLineNo">331</span> + Bytes.toString(key.get(), key.getOffset(), key.getLength()));<a name="line.331"></a> +<span class="sourceLineNo">332</span> }<a name="line.332"></a> +<span class="sourceLineNo">333</span> if (filter == null<a name="line.333"></a> +<span class="sourceLineNo">334</span> || !filter.filterRowKey(KeyValueUtil.createFirstOnRow(key.get(), key.getOffset(),<a name="line.334"></a> +<span class="sourceLineNo">335</span> (short) key.getLength()))) {<a name="line.335"></a> +<span class="sourceLineNo">336</span> processKV(key, result, context, put, delete);<a name="line.336"></a> +<span class="sourceLineNo">337</span> }<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> protected void processKV(ImmutableBytesWritable key, Result result, Context context, Put put,<a name="line.340"></a> +<span class="sourceLineNo">341</span> Delete delete) throws IOException, InterruptedException {<a name="line.341"></a> +<span class="sourceLineNo">342</span> for (Cell kv : result.rawCells()) {<a name="line.342"></a> +<span class="sourceLineNo">343</span> kv = filterKv(filter, kv);<a name="line.343"></a> +<span class="sourceLineNo">344</span> // skip if we filter it out<a name="line.344"></a> +<span class="sourceLineNo">345</span> if (kv == null) continue;<a name="line.345"></a> +<span class="sourceLineNo">346</span><a name="line.346"></a> +<span class="sourceLineNo">347</span> kv = convertKv(kv, cfRenameMap);<a name="line.347"></a> +<span class="sourceLineNo">348</span> // Deletes and Puts are gathered and written when finished<a name="line.348"></a> +<span class="sourceLineNo">349</span> /*<a name="line.349"></a> +<span class="sourceLineNo">350</span> * If there are sequence of mutations and tombstones in an Export, and after Import the same<a name="line.350"></a> +<span class="sourceLineNo">351</span> * sequence should be restored as it is. If we combine all Delete tombstones into single<a name="line.351"></a> +<span class="sourceLineNo">352</span> * request then there is chance of ignoring few DeleteFamily tombstones, because if we<a name="line.352"></a> +<span class="sourceLineNo">353</span> * submit multiple DeleteFamily tombstones in single Delete request then we are maintaining<a name="line.353"></a> +<span class="sourceLineNo">354</span> * only newest in hbase table and ignoring other. Check - HBASE-12065<a name="line.354"></a> +<span class="sourceLineNo">355</span> */<a name="line.355"></a> +<span class="sourceLineNo">356</span> if (CellUtil.isDeleteFamily(kv)) {<a name="line.356"></a> +<span class="sourceLineNo">357</span> Delete deleteFamily = new Delete(key.get());<a name="line.357"></a> +<span class="sourceLineNo">358</span> deleteFamily.addDeleteMarker(kv);<a name="line.358"></a> +<span class="sourceLineNo">359</span> if (durability != null) {<a name="line.359"></a> +<span class="sourceLineNo">360</span> deleteFamily.setDurability(durability);<a name="line.360"></a> +<span class="sourceLineNo">361</span> }<a name="line.361"></a> +<span class="sourceLineNo">362</span> deleteFamily.setClusterIds(clusterIds);<a name="line.362"></a> +<span class="sourceLineNo">363</span> context.write(key, deleteFamily);<a name="line.363"></a> +<span class="sourceLineNo">364</span> } else if (CellUtil.isDelete(kv)) {<a name="line.364"></a> +<span class="sourceLineNo">365</span> if (delete == null) {<a name="line.365"></a> +<span class="sourceLineNo">366</span> delete = new Delete(key.get());<a name="line.366"></a> +<span class="sourceLineNo">367</span> }<a name="line.367"></a> +<span class="sourceLineNo">368</span> delete.addDeleteMarker(kv);<a name="line.368"></a> +<span class="sourceLineNo">369</span> } else {<a name="line.369"></a> +<span class="sourceLineNo">370</span> if (put == null) {<a name="line.370"></a> +<span class="sourceLineNo">371</span> put = new Put(key.get());<a name="line.371"></a> +<span class="sourceLineNo">372</span> }<a name="line.372"></a> +<span class="sourceLineNo">373</span> addPutToKv(put, kv);<a name="line.373"></a> +<span class="sourceLineNo">374</span> }<a name="line.374"></a> +<span class="sourceLineNo">375</span> }<a name="line.375"></a> +<span class="sourceLineNo">376</span> if (put != null) {<a name="line.376"></a> +<span class="sourceLineNo">377</span> if (durability != null) {<a name="line.377"></a> +<span class="sourceLineNo">378</span> put.setDurability(durability);<a name="line.378"></a> +<span class="sourceLineNo">379</span> }<a name="line.379"></a> +<span class="sourceLineNo">380</span> put.setClusterIds(clusterIds);<a name="line.380"></a> +<span class="sourceLineNo">381</span> context.write(key, put);<a name="line.381"></a> +<span class="sourceLineNo">382</span> }<a name="line.382"></a> +<span class="sourceLineNo">383</span> if (delete != null) {<a name="line.383"></a> +<span class="sourceLineNo">384</span> if (durability != null) {<a name="line.384"></a> +<span class="sourceLineNo">385</span> delete.setDurability(durability);<a name="line.385"></a> +<span class="sourceLineNo">386</span> }<a name="line.386"></a> +<span class="sourceLineNo">387</span> delete.setClusterIds(clusterIds);<a name="line.387"></a> +<span class="sourceLineNo">388</span> context.write(key, delete);<a name="line.388"></a> +<span class="sourceLineNo">389</span> }<a name="line.389"></a> <span class="sourceLineNo">390</span> }<a name="line.390"></a> <span class="sourceLineNo">391</span><a name="line.391"></a> -<span class="sourceLineNo">392</span> @Override<a name="line.392"></a> -<span class="sourceLineNo">393</span> public void setup(Context context) {<a name="line.393"></a> -<span class="sourceLineNo">394</span> LOG.info("Setting up " + getClass() + " mapper.");<a name="line.394"></a> -<span class="sourceLineNo">395</span> Configuration conf = context.getConfiguration();<a name="line.395"></a> -<span class="sourceLineNo">396</span> cfRenameMap = createCfRenameMap(conf);<a name="line.396"></a> -<span class="sourceLineNo">397</span> filter = instantiateFilter(conf);<a name="line.397"></a> -<span class="sourceLineNo">398</span> String durabilityStr = conf.get(WAL_DURABILITY);<a name="line.398"></a> -<span class="sourceLineNo">399</span> if(durabilityStr != null){<a name="line.399"></a> -<span class="sourceLineNo">400</span> durability = Durability.valueOf(durabilityStr.toUpperCase());<a name="line.400"></a> -<span class="sourceLineNo">401</span> LOG.info("setting WAL durability to " + durability);<a name="line.401"></a> -<span class="sourceLineNo">402</span> } else {<a name="line.402"></a> -<span class="sourceLineNo">403</span> LOG.info("setting WAL durability to default.");<a name="line.403"></a> -<span class="sourceLineNo">404</span> }<a name="line.404"></a> -<span class="sourceLineNo">405</span> // TODO: This is kind of ugly doing setup of ZKW just to read the clusterid.<a name="line.405"></a> -<span class="sourceLineNo">406</span> ZooKeeperWatcher zkw = null;<a name="line.406"></a> -<span class="sourceLineNo">407</span> Exception ex = null;<a name="line.407"></a> -<span class="sourceLineNo">408</span> try {<a name="line.408"></a> -<span class="sourceLineNo">409</span> zkw = new ZooKeeperWatcher(conf, context.getTaskAttemptID().toString(), null);<a name="line.409"></a> -<span class="sourceLineNo">410</span> clusterIds = Collections.singletonList(ZKClusterId.getUUIDForCluster(zkw));<a name="line.410"></a> -<span class="sourceLineNo">411</span> } catch (ZooKeeperConnectionException e) {<a name="line.411"></a> -<span class="sourceLineNo">412</span> ex = e;<a name="line.412"></a> -<span class="sourceLineNo">413</span> LOG.error("Problem connecting to ZooKeper during task setup", e);<a name="line.413"></a> -<span class="sourceLineNo">414</span> } catch (KeeperException e) {<a name="line.414"></a> -<span class="sourceLineNo">415</span> ex = e;<a name="line.415"></a> -<span class="sourceLineNo">416</span> LOG.error("Problem reading ZooKeeper data during task setup", e);<a name="line.416"></a> -<span class="sourceLineNo">417</span> } catch (IOException e) {<a name="line.417"></a> -<span class="sourceLineNo">418</span> ex = e;<a name="line.418"></a> -<span class="sourceLineNo">419</span> LOG.error("Problem setting up task", e);<a name="line.419"></a> -<span class="sourceLineNo">420</span> } finally {<a name="line.420"></a> -<span class="sourceLineNo">421</span> if (zkw != null) zkw.close();<a name="line.421"></a> -<span class="sourceLineNo">422</span> }<a name="line.422"></a> -<span class="sourceLineNo">423</span> if (clusterIds == null) {<a name="line.423"></a> -<span class="sourceLineNo">424</span> // exit early if setup fails<a name="line.424"></a> -<span class="sourceLineNo">425</span> throw new RuntimeException(ex);<a name="line.425"></a> +<span class="sourceLineNo">392</span> protected void addPutToKv(Put put, Cell kv) throws IOException {<a name="line.392"></a> +<span class="sourceLineNo">393</span> put.add(kv);<a name="line.393"></a> +<span class="sourceLineNo">394</span> }<a name="line.394"></a> +<span class="sourceLineNo">395</span><a name="line.395"></a> +<span class="sourceLineNo">396</span> @Override<a name="line.396"></a> +<span class="sourceLineNo">397</span> public void setup(Context context) {<a name="line.397"></a> +<span class="sourceLineNo">398</span> LOG.info("Setting up " + getClass() + " mapper.");<a name="line.398"></a> +<span class="sourceLineNo">399</span> Configuration conf = context.getConfiguration();<a name="line.399"></a> +<span class="sourceLineNo">400</span> cfRenameMap = createCfRenameMap(conf);<a name="line.400"></a> +<span class="sourceLineNo">401</span> filter = instantiateFilter(conf);<a name="line.401"></a> +<span class="sourceLineNo">402</span> String durabilityStr = conf.get(WAL_DURABILITY);<a name="line.402"></a> +<span class="sourceLineNo">403</span> if(durabilityStr != null){<a name="line.403"></a> +<span class="sourceLineNo">404</span> durability = Durability.valueOf(durabilityStr.toUpperCase());<a name="line.404"></a> +<span class="sourceLineNo">405</span> LOG.info("setting WAL durability to " + durability);<a name="line.405"></a> +<span class="sourceLineNo">406</span> } else {<a name="line.406"></a> +<span class="sourceLineNo">407</span> LOG.info("setting WAL durability to default.");<a name="line.407"></a> +<span class="sourceLineNo">408</span> }<a name="line.408"></a> +<span class="sourceLineNo">409</span> // TODO: This is kind of ugly doing setup of ZKW just to read the clusterid.<a name="line.409"></a> +<span class="sourceLineNo">410</span> ZooKeeperWatcher zkw = null;<a name="line.410"></a> +<span class="sourceLineNo">411</span> Exception ex = null;<a name="line.411"></a> +<span class="sourceLineNo">412</span> try {<a name="line.412"></a> +<span class="sourceLineNo">413</span> zkw = new ZooKeeperWatcher(conf, context.getTaskAttemptID().toString(), null);<a name="line.413"></a> +<span class="sourceLineNo">414</span> clusterIds = Collections.singletonList(ZKClusterId.getUUIDForCluster(zkw));<a name="line.414"></a> +<span class="sourceLineNo">415</span> } catch (ZooKeeperConnectionException e) {<a name="line.415"></a> +<span class="sourceLineNo">416</span> ex = e;<a name="line.416"></a> +<span class="sourceLineNo">417</span> LOG.error("Problem connecting to ZooKeper during task setup", e);<a name="line.417"></a> +<span class="sourceLineNo">418</span> } catch (KeeperException e) {<a name="line.418"></a> +<span class="sourceLineNo">419</span> ex = e;<a name="line.419"></a> +<span class="sourceLineNo">420</span> LOG.error("Problem reading ZooKeeper data during task setup", e);<a name="line.420"></a> +<span class="sourceLineNo">421</span> } catch (IOException e) {<a name="line.421"></a> +<span class="sourceLineNo">422</span> ex = e;<a name="line.422"></a> +<span class="sourceLineNo">423</span> LOG.error("Problem setting up task", e);<a name="line.423"></a> +<span class="sourceLineNo">424</span> } finally {<a name="line.424"></a> +<span class="sourceLineNo">425</span> if (zkw != null) zkw.close();<a name="line.425"></a> <span class="sourceLineNo">426</span> }<a name="line.426"></a> -<span class="sourceLineNo">427</span> }<a name="line.427"></a> -<span class="sourceLineNo">428</span> }<a name="line.428"></a> -<span class="sourceLineNo">429</span><a name="line.429"></a> -<span class="sourceLineNo">430</span> /**<a name="line.430"></a> -<span class="sourceLineNo">431</span> * Create a {@link Filter} to apply to all incoming keys ({@link KeyValue KeyValues}) to<a name="line.431"></a> -<span class="sourceLineNo">432</span> * optionally not include in the job output<a name="line.432"></a> -<span class="sourceLineNo">433</span> * @param conf {@link Configuration} from which to load the filter<a name="line.433"></a> -<span class="sourceLineNo">434</span> * @return the filter to use for the task, or <tt>null</tt> if no filter to should be used<a name="line.434"></a> -<span class="sourceLineNo">435</span> * @throws IllegalArgumentException if the filter is misconfigured<a name="line.435"></a> -<span class="sourceLineNo">436</span> */<a name="line.436"></a> -<span class="sourceLineNo">437</span> public static Filter instantiateFilter(Configuration conf) {<a name="line.437"></a> -<span class="sourceLineNo">438</span> // get the filter, if it was configured <a name="line.438"></a> -<span class="sourceLineNo">439</span> Class<? extends Filter> filterClass = conf.getClass(FILTER_CLASS_CONF_KEY, null, Filter.class);<a name="line.439"></a> -<span class="sourceLineNo">440</span> if (filterClass == null) {<a name="line.440"></a> -<span class="sourceLineNo">441</span> LOG.debug("No configured filter class, accepting all keyvalues.");<a name="line.441"></a> -<span class="sourceLineNo">442</span> return null;<a name="line.442"></a> -<span class="sourceLineNo">443</span> }<a name="line.443"></a> -<span class="sourceLineNo">444</span> LOG.debug("Attempting to create filter:" + filterClass);<a name="line.444"></a> -<span class="sourceLineNo">445</span> String[] filterArgs = conf.getStrings(FILTER_ARGS_CONF_KEY);<a name="line.445"></a> -<span class="sourceLineNo">446</span> ArrayList<byte[]> quotedArgs = toQuotedByteArrays(filterArgs);<a name="line.446"></a> -<span class="sourceLineNo">447</span> try {<a name="line.447"></a> -<span class="sourceLineNo">448</span> Method m = filterClass.getMethod("createFilterFromArguments", ArrayList.class);<a name="line.448"></a> -<span class="sourceLineNo">449</span> return (Filter) m.invoke(null, quotedArgs);<a name="line.449"></a> -<span class="sourceLineNo">450</span> } catch (IllegalAccessException e) {<a name="line.450"></a> -<span class="sourceLineNo">451</span> LOG.error("Couldn't instantiate filter!", e);<a name="line.451"></a> -<span class="sourceLineNo">452</span> throw new RuntimeException(e);<a name="line.452"></a> -<span class="sourceLineNo">453</span> } catch (SecurityException e) {<a name="line.453"></a> -<span class="sourceLineNo">454</span> LOG.error("Couldn't instantiate filter!", e);<a name="line.454"></a> -<span class="sourceLineNo">455</span> throw new RuntimeException(e);<a name="line.455"></a> -<span class="sourceLineNo">456</span> } catch (NoSuchMethodException e) {<a name="line.456"></a> -<span class="sourceLineNo">457</span> LOG.error("Couldn't instantiate filter!", e);<a name="line.457"></a> -<span class="sourceLineNo">458</span> throw new RuntimeException(e);<a name="line.458"></a> -<span class="sourceLineNo">459</span> } catch (IllegalArgumentException e) {<a name="line.459"></a> -<span class="sourceLineNo">460</span> LOG.error("Couldn't instantiate filter!", e);<a name="line.460"></a> -<span class="sourceLineNo">461</span> throw new RuntimeException(e);<a name="line.461"></a> -<span class="sourceLineNo">462</span> } catch (InvocationTargetException e) {<a name="line.462"></a> -<span class="sourceLineNo">463</span> LOG.error("Couldn't instantiate filter!", e);<a name="line.463"></a> -<span class="sourceLineNo">464</span> throw new RuntimeException(e);<a name="line.464"></a> -<span class="sourceLineNo">465</span> }<a name="line.465"></a> -<span class="sourceLineNo">466</span> }<a name="line.466"></a> -<span class="sourceLineNo">467</span><a name="line.467"></a> -<span class="sourceLineNo">468</span> private static ArrayList<byte[]> toQuotedByteArrays(String... stringArgs) {<a name="line.468"></a> -<span class="sourceLineNo">469</span> ArrayList<byte[]> quotedArgs = new ArrayList<byte[]>();<a name="line.469"></a> -<span class="sourceLineNo">470</span> for (String stringArg : stringArgs) {<a name="line.470"></a> -<span class="sourceLineNo">471</span> // all the filters' instantiation methods expected quoted args since they are coming from<a name="line.471"></a> -<span class="sourceLineNo">472</span> // the shell, so add them here, though it shouldn't really be needed :-/<a name="line.472"></a> -<span class="sourceLineNo">473</span> quotedArgs.add(Bytes.toBytes("'" + stringArg + "'"));<a name="line.473"></a> -<span class="sourceLineNo">474</span> }<a name="line.474"></a> -<span class="sourceLineNo">475</span> return quotedArgs;<a name="line.475"></a> -<span class="sourceLineNo">476</span> }<a name="line.476"></a> -<span class="sourceLineNo">477</span><a name="line.477"></a> -<span class="sourceLineNo">478</span> /**<a name="line.478"></a> -<span class="sourceLineNo">479</span> * Attempt to filter out the keyvalue<a name="line.479"></a> -<span class="sourceLineNo">480</span> * @param kv {@link KeyValue} on which to apply the filter<a name="line.480"></a> -<span class="sourceLineNo">481</span> * @return <tt>null</tt> if the key should not be written, otherwise returns the original<a name="line.481"></a> -<span class="sourceLineNo">482</span> * {@link KeyValue}<a name="line.482"></a> -<span class="sourceLineNo">483</span> */<a name="line.483"></a> -<span class="sourceLineNo">484</span> public static Cell filterKv(Filter filter, Cell kv) throws IOException {<a name="line.484"></a> -<span class="sourceLineNo">485</span> // apply the filter and skip this kv if the filter doesn't apply<a name="line.485"></a> -<span class="sourceLineNo">486</span> if (filter != null) {<a name="line.486"></a> -<span class="sourceLineNo">487</span> Filter.ReturnCode code = filter.filterKeyValue(kv);<a name="line.487"></a> -<span class="sourceLineNo">488</span> if (LOG.isTraceEnabled()) {<a name="line.488"></a> -<span class="sourceLineNo">489</span> LOG.trace("Filter returned:" + code + " for the key value:" + kv);<a name="line.489"></a> -<span class="sourceLineNo">490</span> }<a name="line.490"></a> -<span class="sourceLineNo">491</span> // if its not an accept type, then skip this kv<a name="line.491"></a> -<span class="sourceLineNo">492</span> if (!(code.equals(Filter.ReturnCode.INCLUDE) || code<a name="line.492"></a> -<span class="sourceLineNo">493</span> .equals(Filter.ReturnCode.INCLUDE_AND_NEXT_COL))) {<a name="line.493"></a> -<span class="sourceLineNo">494</span> return null;<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> return kv;<a name="line.497"></a> -<span class="sourceLineNo">498</span> }<a name="line.498"></a> -<span class="sourceLineNo">499</span><a name="line.499"></a> -<span class="sourceLineNo">500</span> // helper: create a new KeyValue based on CF rename map<a name="line.500"></a> -<span class="sourceLineNo">501</span> private static Cell convertKv(Cell kv, Map<byte[], byte[]> cfRenameMap) {<a name="line.501"></a> -<span class="sourceLineNo">502</span> if(cfRenameMap != null) {<a name="line.502"></a> -<span class="sourceLineNo">503</span> // If there's a rename mapping for this CF, create a new KeyValue<a name="line.503"></a> -<span class="sourceLineNo">504</span> byte[] newCfName = cfRenameMap.get(CellUtil.cloneFamily(kv));<a name="line.504"></a> -<span class="sourceLineNo">505</span> if(newCfName != null) {<a name="line.505"></a> -<span class="sourceLineNo">506</span> kv = new KeyValue(kv.getRowArray(), // row buffer <a name="line.506"></a> -<span class="sourceLineNo">507</span> kv.getRowOffset(), // row offset<a name="line.507"></a> -<span class="sourceLineNo">508</span> kv.getRowLength(), // row length<a name="line.508"></a> -<span class="sourceLineNo">509</span> newCfName, // CF buffer<a name="line.509"></a> -<span class="sourceLineNo">510</span> 0, // CF offset <a name="line.510"></a> -<span class="sourceLineNo">511</span> newCfName.length, // CF length <a name="line.511"></a> -<span class="sourceLineNo">512</span> kv.getQualifierArray(), // qualifier buffer<a name="line.512"></a> -<span class="sourceLineNo">513</span> kv.getQualifierOffset(), // qualifier offset<a name="line.513"></a> -<span class="sourceLineNo">514</span> kv.getQualifierLength(), // qualifier length<a name="line.514"></a> -<span class="sourceLineNo">515</span> kv.getTimestamp(), // timestamp<a name="line.515"></a> -<span class="sourceLineNo">516</span> KeyValue.Type.codeToType(kv.getTypeByte()), // KV Type<a name="line.516"></a> -<span class="sourceLineNo">517</span> kv.getValueArray(), // value buffer <a name="line.517"></a> -<span class="sourceLineNo">518</span> kv.getValueOffset(), // value offset<a name="line.518"></a> -<span class="sourceLineNo">519</span> kv.getValueLength()); // value length<a name="line.519"></a> -<span class="sourceLineNo">520</span> }<a name="line.520"></a> -<span class="sourceLineNo">521</span> }<a name="line.521"></a> -<span class="sourceLineNo">522</span> return kv;<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> // helper: make a map from sourceCfName to destCfName by parsing a config key<a name="line.525"></a> -<span class="sourceLineNo">526</span> private static Map<byte[], byte[]> createCfRenameMap(Configuration conf) {<a name="line.526"></a> -<span class="sourceLineNo">527</span> Map<byte[], byte[]> cfRenameMap = null;<a name="line.527"></a> -<span class="sourceLineNo">528</span> String allMappingsPropVal = conf.get(CF_RENAME_PROP);<a name="line.528"></a> -<span class="sourceLineNo">529</span> if(allMappingsPropVal != null) {<a name="line.529"></a> -<span class="sourceLineNo">530</span> // The conf value format should be sourceCf1:destCf1,sourceCf2:destCf2,...<a name="line.530"></a> -<span class="sourceLineNo">531</span> String[] allMappings = allMappingsPropVal.split(",");<a name="line.531"></a> -<span class="sourceLineNo">532</span> for (String mapping: allMappings) {<a name="line.532"></a> -<span class="sourceLineNo">533</span> if(cfRenameMap == null) {<a name="line.533"></a> -<span class="sourceLineNo">534</span> cfRenameMap = new TreeMap<byte[],byte[]>(Bytes.BYTES_COMPARATOR);<a name="line.534"></a> -<span class="sourceLineNo">535</span> }<a name="line.535"></a> -<span class="sourceLineNo">536</span> String [] srcAndDest = mapping.split(":");<a name="line.536"></a> -<span class="sourceLineNo">537</span> if(srcAndDest.length != 2) {<a name="line.537"></a> -<span class="sourceLineNo">538</span> continue;<a name="line.538"></a> +<span class="sourceLineNo">427</span> if (clusterIds == null) {<a name="line.427"></a> +<span class="sourceLineNo">428</span> // exit early if setup fails<a name="line.428"></a> +<span class="sourceLineNo">429</span> throw new RuntimeException(ex);<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> /**<a name="line.434"></a> +<span class="sourceLineNo">435</span> * Create a {@link Filter} to apply to all incoming keys ({@link KeyValue KeyValues}) to<a name="line.435"></a> +<span class="sourceLineNo">436</span> * optionally not include in the job output<a name="line.436"></a> +<span class="sourceLineNo">437</span> * @param conf {@link Configuration} from which to load the filter<a name="line.437"></a> +<span class="sourceLineNo">438</span> * @return the filter to use for the task, or <tt>null</tt> if no filter to should be used<a name="line.438"></a> +<span class="sourceLineNo">439</span> * @throws IllegalArgumentException if the filter is misconfigured<a name="line.439"></a> +<span class="sourceLineNo">440</span> */<a name="line.440"></a> +<span class="sourceLineNo">441</span> public static Filter instantiateFilter(Configuration conf) {<a name="line.441"></a> +<span class="sourceLineNo">442</span> // get the filter, if it was configured <a name="line.442"></a> +<span class="sourceLineNo">443</span> Class<? extends Filter> filterClass = conf.getClass(FILTER_CLASS_CONF_KEY, null, Filter.class);<a name="line.443"></a> +<span class="sourceLineNo">444</span> if (filterClass == null) {<a name="line.444"></a> +<span class="sourceLineNo">445</span> LOG.debug("No configured filter class, accepting all keyvalues.");<a name="line.445"></a> +<span class="sourceLineNo">446</span> return null;<a name="line.446"></a> +<span class="sourceLineNo">447</span> }<a name="line.447"></a> +<span class="sourceLineNo">448</span> LOG.debug("Attempting to create filter:" + filterClass);<a name="line.448"></a> +<span class="sourceLineNo">449</span> String[] filterArgs = conf.getStrings(FILTER_ARGS_CONF_KEY);<a name="line.449"></a> +<span class="sourceLineNo">450</span> ArrayList<byte[]> quotedArgs = toQuotedByteArrays(filterArgs);<a name="line.450"></a> +<span class="sourceLineNo">451</span> try {<a name="line.451"></a> +<span class="sourceLineNo">452</span> Method m = filterClass.getMethod("createFilterFromArguments", ArrayList.class);<a name="line.452"></a> +<span class="sourceLineNo">453</span> return (Filter) m.invoke(null, quotedArgs);<a name="line.453"></a> +<span class="sourceLineNo">454</span> } catch (IllegalAccessException e) {<a name="line.454"></a> +<span class="sourceLineNo">455</span> LOG.error("Couldn't instantiate filter!", e);<a name="line.455"></a> +<span class="sourceLineNo">456</span> throw new RuntimeException(e);<a name="line.456"></a> +<span class="sourceLineNo">457</span> } catch (SecurityException e) {<a name="line.457"></a> +<span class="sourceLineNo">458</span> LOG.error("Couldn't instantiate filter!", e);<a name="line.458"></a> +<span class="sourceLineNo">459</span> throw new RuntimeException(e);<a name="line.459"></a> +<span class="sourceLineNo">460</span> } catch (NoSuchMethodException e) {<a name="line.460"></a> +<span class="sourceLineNo">461</span> LOG.error("Couldn't instantiate filter!", e);<a name="line.461"></a> +<span class="sourceLineNo">462</span> throw new RuntimeException(e);<a name="line.462"></a> +<span class="sourceLineNo">463</span> } catch (IllegalArgumentException e) {<a name="line.463"></a> +<span class="sourceLineNo">464</span> LOG.error("Couldn't instantiate filter!", e);<a name="line.464"></a> +<span class="sourceLineNo">465</span> throw new RuntimeException(e);<a name="line.465"></a> +<span class="sourceLineNo">466</span> } catch (InvocationTargetException e) {<a name="line.466"></a> +<span class="sourceLineNo">467</span> LOG.error("Couldn't instantiate filter!", e);<a name="line.467"></a> +<span class="sourceLineNo">468</span> throw new RuntimeException(e);<a name="line.468"></a> +<span class="sourceLineNo">469</span> }<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> private static ArrayList<byte[]> toQuotedByteArrays(String... stringArgs) {<a name="line.472"></a> +<span class="sourceLineNo">473</span> ArrayList<byte[]> quotedArgs = new ArrayList<byte[]>();<a name="line.473"></a> +<span class="sourceLineNo">474</span> for (String stringArg : stringArgs) {<a name="line.474"></a> +<span class="sourceLineNo">475</span> // all the filters' instantiation methods expected quoted args since they are coming from<a name="line.475"></a> +<span class="sourceLineNo">476</span> // the shell, so add them here, though it shouldn't really be needed :-/<a name="line.476"></a> +<span class="sourceLineNo">477</span> quotedArgs.add(Bytes.toBytes("'" + stringArg + "'"));<a name="line.477"></a> +<span class="sourceLineNo">478</span> }<a name="line.478"></a> +<span class="sourceLineNo">479</span> return quotedArgs;<a name="line.479"></a> +<span class="sourceLineNo">480</span> }<a name="line.480"></a> +<span class="sourceLineNo">481</span><a name="line.481"></a> +<span class="sourceLineNo">482</span> /**<a name="line.482"></a> +<span class="sourceLineNo">483</span> * Attempt to filter out the keyvalue<a name="line.483"></a> +<span class="sourceLineNo">484</span> * @param kv {@link KeyValue} on which to apply the filter<a name="line.484"></a> +<span class="sourceLineNo">485</span> * @return <tt>null</tt> if the key should not be written, otherwise returns the original<a name="line.485"></a> +<span class="sourceLineNo">486</span> * {@link KeyValue}<a name="line.486"></a> +<span class="sourceLineNo">487</span> */<a name="line.487"></a> +<span class="sourceLineNo">488</span> public static Cell filterKv(Filter filter, Cell kv) throws IOException {<a name="line.488"></a> +<span class="sourceLineNo">489</span> // apply the filter and skip this kv if the filter doesn't apply<a name="line.489"></a> +<span class="sourceLineNo">490</span> if (filter != null) {<a name="line.490"></a> +<span class="sourceLineNo">491</span> Filter.ReturnCode code = filter.filterKeyValue(kv);<a name="line.491"></a> +<span class="sourceLineNo">492</span> if (LOG.isTraceEnabled()) {<a name="line.492"></a> +<span class="sourceLineNo">493</span> LOG.trace("Filter returned:" + code + " for the key value:" + kv);<a name="line.493"></a> +<span class="sourceLineNo">494</span> }<a name="line.494"></a> +<span class="sourceLineNo">495</span> // if its not an accept type, then skip this kv<a name="line.495"></a> +<span class="sourceLineNo">496</span> if (!(code.equals(Filter.ReturnCode.INCLUDE) || code<a name="line.496"></a> +<span class="sourceLineNo">497</span> .equals(Filter.ReturnCode.INCLUDE_AND_NEXT_COL))) {<a name="line.497"></a> +<span class="sourceLineNo">498</span> return null;<a name="line.498"></a> +<span class="sourceLineNo">499</span> }<a name="line.499"></a> +<span class="sourceLineNo">500</span> }<a name="line.500"></a> +<span class="sourceLineNo">501</span> return kv;<a name="line.501"></a> +<span class="sourceLineNo">502</span> }<a name="line.502"></a> +<span class="sourceLineNo">503</span><a name="line.503"></a> +<span class="sourceLineNo">504</span> // helper: create a new KeyValue based on CF rename map<a name="line.504"></a> +<span class="sourceLineNo">505</span> private static Cell convertKv(Cell kv, Map<byte[], byte[]> cfRenameMap) {<a name="line.505"></a> +<span class="sourceLineNo">506</span> if(cfRenameMap != null) {<a name="line.506"></a> +<span class="sourceLineNo">507</span> // If there's a rename mapping for this CF, create a new KeyValue<a name="line.507"></a> +<span class="sourceLineNo">508</span> byte[] newCfName = cfRenameMap.get(CellUtil.cloneFamily(kv));<a name="line.508"></a> +<span class="sourceLineNo">509</span> if(newCfName != null) {<a name="line.509"></a> +<span class="sourceLineNo">510</span> kv = new KeyValue(kv.getRowArray(), // row buffer <a name="line.510"></a> +<span class="sourceLineNo">511</span> kv.getRowOffset(), // row offset<a name="line.511"></a> +<span class="sourceLineNo">512</span> kv.getRowLength(), // row length<a name="line.512"></a> +<span class="sourceLineNo">513</span> newCfName, // CF buffer<a name="line.513"></a> +<span class="sourceLineNo">514</span> 0, // CF offset <a name="line.514"></a> +<span class="sourceLineNo">515</span> newCfName.length, // CF length <a name="line.515"></a> +<span class="sourceLineNo">516</span> kv.getQualifierArray(), // qualifier buffer<a name="line.516"></a> +<span class="sourceLineNo">517</span> kv.getQualifierOffset(), // qualifier offset<a name="line.517"></a> +<span class="sourceLineNo">518</span> kv.getQualifierLength(), // qualifier length<a name="line.518"></a> +<span class="sourceLineNo">519</span> kv.getTimestamp(), // timestamp<a name="line.519"></a> +<span class="sourceLineNo">520</span> KeyValue.Type.codeToType(kv.getTypeByte()), // KV Type<a name="line.520"></a> +<span class="sourceLineNo">521</span> kv.getValueArray(), // value buffer <a name="line.521"></a> +<span class="sourceLineNo">522</span> kv.getValueOffset(), // value offset<a name="line.522"></a> +<span class="sourceLineNo">523</span> kv.getValueLength()); // value length<a name="line.523"></a> +<span class="sourceLineNo">524</span> }<a name="line.524"></a> +<span class="sourceLineNo">525</span> }<a name="line.525"></a> +<span class="sourceLineNo">526</span> return kv;<a name="line.526"></a> +<span class="sourceLineNo">527</span> }<a name="line.527"></a> +<span class="sourceLineNo">528</span><a name="line.528"></a> +<span class="sourceLineNo">529</span> // helper: make a map from sourceCfName to destCfName by parsing a config key<a name="line.529"></a> +<span class="sourceLineNo">530</span> private static Map<byte[], byte[]> createCfRenameMap(Configuration conf) {<a name="line.530"></a> +<span class="sourceLineNo">531</span> Map<byte[], byte[]> cfRenameMap = null;<a name="line.531"></a> +<span class="sourceLineNo">532</span> String allMappingsPropVal = conf.get(CF_RENAME_PROP);<a name="line.532"></a> +<span class="sourceLineNo">533</span> if(allMappingsPropVal != null) {<a name="line.533"></a> +<span class="sourceLineNo">534</span> // The conf value format should be sourceCf1:destCf1,sourceCf2:destCf2,...<a name="line.534"></a> +<span class="sourceLineNo">535</span> String[] allMappings = allMappingsPropVal.split(",");<a name="line.535"></a> +<span class="sourceLineNo">536</span> for (String mapping: allMappings) {<a name="line.536"></a> +<span class="sourceLineNo">537</span> if(cfRenameMap == null) {<a name="line.537"></a> +<span class="sourceLineNo">538</span> cfRenameMap = new TreeMap<byte[],byte[]>(Bytes.BYTES_COMPARATOR);<a name="line.538"></a> <span class="sourceLineNo">539</span> }<a name="line.539"></a> -<span class="sourceLineNo">540</span> cfRenameMap.put(srcAndDest[0].getBytes(), srcAndDest[1].getBytes());<a name="line.540"></a> -<span class="sourceLineNo">541</span> }<a name="line.541"></a> -<span class="sourceLineNo">542</span> }<a name="line.542"></a> -<span class="sourceLineNo">543</span> return cfRenameMap;<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> /**<a name="line.546"></a> -<span class="sourceLineNo">547</span> * <p>Sets a configuration property with key {@link #CF_RENAME_PROP} in conf that tells<a name="line.547"></a> -<span class="sourceLineNo">548</span> * the mapper how to rename column families.<a name="line.548"></a> -<span class="sourceLineNo">549</span> * <a name="line.549"></a> -<span class="sourceLineNo">550</span> * <p>Alternately, instead of calling this function, you could set the configuration key <a name="line.550"></a> -<span class="sourceLineNo">551</span> * {@link #CF_RENAME_PROP} yourself. The value should look like <a name="line.551"></a> -<span class="sourceLineNo">552</span> * <pre>srcCf1:destCf1,srcCf2:destCf2,....</pre>. This would have the same effect on<a name="line.552"></a> -<span class="sourceLineNo">553</span> * the mapper behavior.<a name="line.553"></a> -<span class="sourceLineNo">554</span> * <a name="line.554"></a> -<span class="sourceLineNo">555</span> * @param conf the Configuration in which the {@link #CF_RENAME_PROP} key will be<a name="line.555"></a> -<span class="sourceLineNo">556</span> * set<a name="line.556"></a> -<span class="sourceLineNo">557</span> * @param renameMap a mapping from source CF names to destination CF names<a name="line.557"></a> -<span class="sourceLineNo">558</span> */<a name="line.558"></a> -<span class="sourceLineNo">559</span> static public void configureCfRenaming(Configuration conf, <a name="line.559"></a> -<span class="sourceLineNo">560</span> Map<String, String> renameMap) {<a name="line.560"></a> -<span class="sourceLineNo">561</span> StringBuilder sb = new StringBuilder();<a name="line.561"></a> -<span class="sourceLineNo">562</span> for(Map.Entry<String,String> entry: renameMap.entrySet()) {<a name="line.562"></a> -<span class="sourceLineNo">563</span> String sourceCf = entry.getKey();<a name="line.563"></a> -<span class="sourceLineNo">564</span> String destCf = entry.getValue();<a name="line.564"></a> -<span class="sourceLineNo">565</span><a name="line.565"></a> -<span class="sourceLineNo">566</span> if(sourceCf.contains(":") || sourceCf.contains(",") || <a name="line.566"></a> -<span class="sourceLineNo">567</span> destCf.contains(":") || destCf.contains(",")) {<a name="line.567"></a> -<span class="sourceLineNo">568</span> throw new IllegalArgumentException("Illegal character in CF names: " <a name="line.568"></a> -<span class="sourceLineNo">569</span> + sourceCf + ", " + destCf);<a name="line.569"></a> -<span class="sourceLineNo">570</span> }<a name="line.570"></a> -<span class="sourceLineNo">571</span><a name="line.571"></a> -<span class="sourceLineNo">572</span> if(sb.length() != 0) {<a name="line.572"></a> -<span class="sourceLineNo">573</span> sb.append(",");<a name="line.573"></a> +<span class="sourceLineNo">540</span> String [] srcAndDest = mapping.split(":");<a name="line.540"></a> +<span class="sourceLineNo">541</span> if(srcAndDest.length != 2) {<a name="line.541"></a> +<span class="sourceLineNo">542</span> continue;<a name="line.542"></a> +<span class="sourceLineNo">543</span> }<a name="line.543"></a> +<span class="sourceLineNo">544</span> cfRenameMap.put(srcAndDest[0].getBytes(), srcAndDest[1].getBytes());<a name="line.544"></a> +<span class="sourceLineNo">545</span> }<a name="line.545"></a> +<span class="sourceLineNo">546</span> }<a name="line.546"></a> +<span class="sourceLineNo">547</span> return cfRenameMap;<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> * <p>Sets a configuration property with key {@link #CF_RENAME_PROP} in conf that tells<a name="line.551"></a> +<span class="sourceLineNo">552</span> * the mapper how to rename column families.<a name="line.552"></a> +<span class="sourceLineNo">553</span> * <a name="line.553"></a> +<span class="sourceLineNo">554</span> * <p>Alternately, instead of calling this function, you could set the configuration key <a name="line.554"></a> +<span class="sourceLineNo">555</span> * {@link #CF_RENAME_PROP} yourself. The value should look like <a name="line.555"></a> +<span class="sourceLineNo">556</span> * <pre>srcCf1:destCf1,srcCf2:destCf2,....</pre>. This would have the same effect on<a name="line.556"></a> +<span class="sourceLineNo">557</span> * the mapper behavior.<a name="line.557"></a> +<span class="sourceLineNo">558</span> * <a name="line.558"></a> +<span class="sourceLineNo">559</span> * @param conf the Configuration in which the {@link #CF_RENAME_PROP} key will be<a name="line.559"></a> +<span class="sourceLineNo">560</span> * set<a name="line.560"></a> +<span class="sourceLineNo">561</span> * @param renameMap a mapping from source CF names to destination CF names<a name="line.561"></a> +<span class="sourceLineNo">562</span> */<a name="line.562"></a> +<span class="sourceLineNo">563</span> static public void configureCfRenaming(Configuration conf, <a name="line.563"></a> +<span class="sourceLineNo">564</span> Map<String, String> renameMap) {<a name="line.564"></a> +<span class="sourceLineNo">565</span> StringBuilder sb = new StringBuilder();<a name="line.565"></a> +<span class="sourceLineNo">566</span> for(Map.Entry<String,String> entry: renameMap.entrySet()) {<a name="line.566"></a> +<span class="sourceLineNo">567</span> String sourceCf = entry.getKey();<a name="line.567"></a> +<span class="sourceLineNo">568</span> String destCf = entry.getValue();<a name="line.568"></a> +<span class="sourceLineNo">569</span><a name="line.569"></a> +<span class="sourceLineNo">570</span> if(sourceCf.contains(":") || sourceCf.contains(",") || <a name="line.570"></a> +<span class="sourceLineNo">571</span> destCf.contains(":") || destCf.contains(",")) {<a name="line.571"></a> +<span class="sourceLineNo">572</span> throw new IllegalArgumentException("Illegal character in CF names: " <a name="line.572"></a> +<span class="sourceLineNo">573</span> + sourceCf + ", " + destCf);<a name="line.573"></a> <span class="sourceLineNo">574</span> }<a name="line.574"></a> -<span class="sourceLineNo">575</span> sb.append(sourceCf + ":" + destCf);<a name="line.575"></a> -<span class="sourceLineNo">576</span> }<a name="line.576"></a> -<span class="sourceLineNo">577</span> conf.set(CF_RENAME_PROP, sb.toString());<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> -<span class="sourceLineNo">580</span> /**<a name="line.580"></a> -<span class="sourceLineNo">581</span> * Add a Filter to be instantiated on import<a name="line.581"></a> -<span class="sourceLineNo">582</span> * @param conf Configuration to update (will be passed to the job)<a name="line.582"></a> -<span class="sourceLineNo">583</span> * @param clazz {@link Filter} subclass to instantiate on the server.<a name="line.583"></a> -<span class="sourceLineNo">584</span> * @param filterArgs List of arguments to pass to the filter on instantiation<a name="line.584"></a> -<span class="sourceLineNo">585</span> */<a name="line.585"></a> -<span class="sourceLineNo">586</span> public static void addFilterAndArguments(Configuration conf, Class<? extends Filter> clazz,<a name="line.586"></a> -<span class="sourceLineNo">587</span> List<String> filterArgs) throws IOException {<a name="line.587"></a> -<span class="sourceLineNo">588</span> conf.set(Import.FILTER_CLASS_CONF_KEY, clazz.getName());<a name="line.588"></a> -<span class="sourceLineNo">589</span> conf.setStrings(Import.FILTER_ARGS_CONF_KEY, filterArgs.toArray(new String[filterArgs.size()]));<a name="line.589"></a> -<span class="sourceLineNo">590</span> }<a name="line.590"></a> -<span class="sourceLineNo">591</span><a name="line.591"></a> -<span class="sourceLineNo">592</span> /**<a name="line.592"></a> -<span class="sourceLineNo">593</span> * Sets up the actual job.<a name="line.593"></a> -<span class="sourceLineNo">594</span> * @param conf The current configuration.<a name="line.594"></a> -<span class="sourceLineNo">595</span> * @param args The command line parameters.<a name="line.595"></a> -<span class="sourceLineNo">596</span> * @return The newly created job.<a name="line.596"></a> -<span class="sourceLineNo">597</span> * @throws IOException When setting up the job fails.<a name="line.597"></a> -<span class="sourceLineNo
<TRUNCATED>