1 | /** |
---|
2 | * Licensed to the Apache Software Foundation (ASF) under one or more |
---|
3 | * contributor license agreements. See the NOTICE file distributed with this |
---|
4 | * work for additional information regarding copyright ownership. The ASF |
---|
5 | * licenses this file to you under the Apache License, Version 2.0 (the |
---|
6 | * "License"); you may not use this file except in compliance with the License. |
---|
7 | * You may obtain a copy of the License at |
---|
8 | * |
---|
9 | * http://www.apache.org/licenses/LICENSE-2.0 |
---|
10 | * |
---|
11 | * Unless required by applicable law or agreed to in writing, software |
---|
12 | * distributed under the License is distributed on an "AS IS" BASIS, WITHOUT |
---|
13 | * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the |
---|
14 | * License for the specific language governing permissions and limitations under |
---|
15 | * the License. |
---|
16 | */ |
---|
17 | |
---|
18 | package org.apache.hadoop.io.file.tfile; |
---|
19 | |
---|
20 | import java.io.DataInputStream; |
---|
21 | import java.io.DataOutputStream; |
---|
22 | import java.io.IOException; |
---|
23 | import java.util.Arrays; |
---|
24 | |
---|
25 | import junit.framework.TestCase; |
---|
26 | |
---|
27 | import org.apache.hadoop.conf.Configuration; |
---|
28 | import org.apache.hadoop.fs.FSDataInputStream; |
---|
29 | import org.apache.hadoop.fs.FSDataOutputStream; |
---|
30 | import org.apache.hadoop.fs.FileSystem; |
---|
31 | import org.apache.hadoop.fs.Path; |
---|
32 | import org.apache.hadoop.io.file.tfile.TFile.Reader; |
---|
33 | import org.apache.hadoop.io.file.tfile.TFile.Writer; |
---|
34 | import org.apache.hadoop.io.file.tfile.TFile.Reader.Scanner; |
---|
35 | |
---|
36 | /** |
---|
37 | * test tfile features. |
---|
38 | * |
---|
39 | */ |
---|
40 | public class TestTFile extends TestCase { |
---|
41 | private static String ROOT = |
---|
42 | System.getProperty("test.build.data", "/tmp/tfile-test"); |
---|
43 | private FileSystem fs; |
---|
44 | private Configuration conf; |
---|
45 | private final int minBlockSize = 512; |
---|
46 | private final int largeVal = 3 * 1024 * 1024; |
---|
47 | private static String localFormatter = "%010d"; |
---|
48 | |
---|
49 | @Override |
---|
50 | public void setUp() throws IOException { |
---|
51 | conf = new Configuration(); |
---|
52 | fs = FileSystem.get(conf); |
---|
53 | } |
---|
54 | |
---|
55 | @Override |
---|
56 | public void tearDown() throws IOException { |
---|
57 | // do nothing |
---|
58 | } |
---|
59 | |
---|
60 | // read a key from the scanner |
---|
61 | public byte[] readKey(Scanner scanner) throws IOException { |
---|
62 | int keylen = scanner.entry().getKeyLength(); |
---|
63 | byte[] read = new byte[keylen]; |
---|
64 | scanner.entry().getKey(read); |
---|
65 | return read; |
---|
66 | } |
---|
67 | |
---|
68 | // read a value from the scanner |
---|
69 | public byte[] readValue(Scanner scanner) throws IOException { |
---|
70 | int valueLen = scanner.entry().getValueLength(); |
---|
71 | byte[] read = new byte[valueLen]; |
---|
72 | scanner.entry().getValue(read); |
---|
73 | return read; |
---|
74 | } |
---|
75 | |
---|
76 | // read a long value from the scanner |
---|
77 | public byte[] readLongValue(Scanner scanner, int len) throws IOException { |
---|
78 | DataInputStream din = scanner.entry().getValueStream(); |
---|
79 | byte[] b = new byte[len]; |
---|
80 | din.readFully(b); |
---|
81 | din.close(); |
---|
82 | return b; |
---|
83 | } |
---|
84 | |
---|
85 | // write some records into the tfile |
---|
86 | // write them twice |
---|
87 | private int writeSomeRecords(Writer writer, int start, int n) |
---|
88 | throws IOException { |
---|
89 | String value = "value"; |
---|
90 | for (int i = start; i < (start + n); i++) { |
---|
91 | String key = String.format(localFormatter, i); |
---|
92 | writer.append(key.getBytes(), (value + key).getBytes()); |
---|
93 | writer.append(key.getBytes(), (value + key).getBytes()); |
---|
94 | } |
---|
95 | return (start + n); |
---|
96 | } |
---|
97 | |
---|
98 | // read the records and check |
---|
99 | private int readAndCheckbytes(Scanner scanner, int start, int n) |
---|
100 | throws IOException { |
---|
101 | String value = "value"; |
---|
102 | for (int i = start; i < (start + n); i++) { |
---|
103 | byte[] key = readKey(scanner); |
---|
104 | byte[] val = readValue(scanner); |
---|
105 | String keyStr = String.format(localFormatter, i); |
---|
106 | String valStr = value + keyStr; |
---|
107 | assertTrue("btyes for keys do not match " + keyStr + " " |
---|
108 | + new String(key), Arrays.equals(keyStr.getBytes(), key)); |
---|
109 | assertTrue("bytes for vals do not match " + valStr + " " |
---|
110 | + new String(val), Arrays.equals( |
---|
111 | valStr.getBytes(), val)); |
---|
112 | assertTrue(scanner.advance()); |
---|
113 | key = readKey(scanner); |
---|
114 | val = readValue(scanner); |
---|
115 | assertTrue("btyes for keys do not match", Arrays.equals( |
---|
116 | keyStr.getBytes(), key)); |
---|
117 | assertTrue("bytes for vals do not match", Arrays.equals( |
---|
118 | valStr.getBytes(), val)); |
---|
119 | assertTrue(scanner.advance()); |
---|
120 | } |
---|
121 | return (start + n); |
---|
122 | } |
---|
123 | |
---|
124 | // write some large records |
---|
125 | // write them twice |
---|
126 | private int writeLargeRecords(Writer writer, int start, int n) |
---|
127 | throws IOException { |
---|
128 | byte[] value = new byte[largeVal]; |
---|
129 | for (int i = start; i < (start + n); i++) { |
---|
130 | String key = String.format(localFormatter, i); |
---|
131 | writer.append(key.getBytes(), value); |
---|
132 | writer.append(key.getBytes(), value); |
---|
133 | } |
---|
134 | return (start + n); |
---|
135 | } |
---|
136 | |
---|
137 | // read large records |
---|
138 | // read them twice since its duplicated |
---|
139 | private int readLargeRecords(Scanner scanner, int start, int n) |
---|
140 | throws IOException { |
---|
141 | for (int i = start; i < (start + n); i++) { |
---|
142 | byte[] key = readKey(scanner); |
---|
143 | String keyStr = String.format(localFormatter, i); |
---|
144 | assertTrue("btyes for keys do not match", Arrays.equals( |
---|
145 | keyStr.getBytes(), key)); |
---|
146 | scanner.advance(); |
---|
147 | key = readKey(scanner); |
---|
148 | assertTrue("btyes for keys do not match", Arrays.equals( |
---|
149 | keyStr.getBytes(), key)); |
---|
150 | scanner.advance(); |
---|
151 | } |
---|
152 | return (start + n); |
---|
153 | } |
---|
154 | |
---|
155 | // write empty keys and values |
---|
156 | private void writeEmptyRecords(Writer writer, int n) throws IOException { |
---|
157 | byte[] key = new byte[0]; |
---|
158 | byte[] value = new byte[0]; |
---|
159 | for (int i = 0; i < n; i++) { |
---|
160 | writer.append(key, value); |
---|
161 | } |
---|
162 | } |
---|
163 | |
---|
164 | // read empty keys and values |
---|
165 | private void readEmptyRecords(Scanner scanner, int n) throws IOException { |
---|
166 | byte[] key = new byte[0]; |
---|
167 | byte[] value = new byte[0]; |
---|
168 | byte[] readKey = null; |
---|
169 | byte[] readValue = null; |
---|
170 | for (int i = 0; i < n; i++) { |
---|
171 | readKey = readKey(scanner); |
---|
172 | readValue = readValue(scanner); |
---|
173 | assertTrue("failed to match keys", Arrays.equals(readKey, key)); |
---|
174 | assertTrue("failed to match values", Arrays.equals(readValue, value)); |
---|
175 | assertTrue("failed to advance cursor", scanner.advance()); |
---|
176 | } |
---|
177 | } |
---|
178 | |
---|
179 | private int writePrepWithKnownLength(Writer writer, int start, int n) |
---|
180 | throws IOException { |
---|
181 | // get the length of the key |
---|
182 | String key = String.format(localFormatter, start); |
---|
183 | int keyLen = key.getBytes().length; |
---|
184 | String value = "value" + key; |
---|
185 | int valueLen = value.getBytes().length; |
---|
186 | for (int i = start; i < (start + n); i++) { |
---|
187 | DataOutputStream out = writer.prepareAppendKey(keyLen); |
---|
188 | String localKey = String.format(localFormatter, i); |
---|
189 | out.write(localKey.getBytes()); |
---|
190 | out.close(); |
---|
191 | out = writer.prepareAppendValue(valueLen); |
---|
192 | String localValue = "value" + localKey; |
---|
193 | out.write(localValue.getBytes()); |
---|
194 | out.close(); |
---|
195 | } |
---|
196 | return (start + n); |
---|
197 | } |
---|
198 | |
---|
199 | private int readPrepWithKnownLength(Scanner scanner, int start, int n) |
---|
200 | throws IOException { |
---|
201 | for (int i = start; i < (start + n); i++) { |
---|
202 | String key = String.format(localFormatter, i); |
---|
203 | byte[] read = readKey(scanner); |
---|
204 | assertTrue("keys not equal", Arrays.equals(key.getBytes(), read)); |
---|
205 | String value = "value" + key; |
---|
206 | read = readValue(scanner); |
---|
207 | assertTrue("values not equal", Arrays.equals(value.getBytes(), read)); |
---|
208 | scanner.advance(); |
---|
209 | } |
---|
210 | return (start + n); |
---|
211 | } |
---|
212 | |
---|
213 | private int writePrepWithUnkownLength(Writer writer, int start, int n) |
---|
214 | throws IOException { |
---|
215 | for (int i = start; i < (start + n); i++) { |
---|
216 | DataOutputStream out = writer.prepareAppendKey(-1); |
---|
217 | String localKey = String.format(localFormatter, i); |
---|
218 | out.write(localKey.getBytes()); |
---|
219 | out.close(); |
---|
220 | String value = "value" + localKey; |
---|
221 | out = writer.prepareAppendValue(-1); |
---|
222 | out.write(value.getBytes()); |
---|
223 | out.close(); |
---|
224 | } |
---|
225 | return (start + n); |
---|
226 | } |
---|
227 | |
---|
228 | private int readPrepWithUnknownLength(Scanner scanner, int start, int n) |
---|
229 | throws IOException { |
---|
230 | for (int i = start; i < start; i++) { |
---|
231 | String key = String.format(localFormatter, i); |
---|
232 | byte[] read = readKey(scanner); |
---|
233 | assertTrue("keys not equal", Arrays.equals(key.getBytes(), read)); |
---|
234 | try { |
---|
235 | read = readValue(scanner); |
---|
236 | assertTrue(false); |
---|
237 | } |
---|
238 | catch (IOException ie) { |
---|
239 | // should have thrown exception |
---|
240 | } |
---|
241 | String value = "value" + key; |
---|
242 | read = readLongValue(scanner, value.getBytes().length); |
---|
243 | assertTrue("values nto equal", Arrays.equals(read, value.getBytes())); |
---|
244 | scanner.advance(); |
---|
245 | } |
---|
246 | return (start + n); |
---|
247 | } |
---|
248 | |
---|
249 | private byte[] getSomeKey(int rowId) { |
---|
250 | return String.format(localFormatter, rowId).getBytes(); |
---|
251 | } |
---|
252 | |
---|
253 | private void writeRecords(Writer writer) throws IOException { |
---|
254 | writeEmptyRecords(writer, 10); |
---|
255 | int ret = writeSomeRecords(writer, 0, 100); |
---|
256 | ret = writeLargeRecords(writer, ret, 1); |
---|
257 | ret = writePrepWithKnownLength(writer, ret, 40); |
---|
258 | ret = writePrepWithUnkownLength(writer, ret, 50); |
---|
259 | writer.close(); |
---|
260 | } |
---|
261 | |
---|
262 | private void readAllRecords(Scanner scanner) throws IOException { |
---|
263 | readEmptyRecords(scanner, 10); |
---|
264 | int ret = readAndCheckbytes(scanner, 0, 100); |
---|
265 | ret = readLargeRecords(scanner, ret, 1); |
---|
266 | ret = readPrepWithKnownLength(scanner, ret, 40); |
---|
267 | ret = readPrepWithUnknownLength(scanner, ret, 50); |
---|
268 | } |
---|
269 | |
---|
270 | private FSDataOutputStream createFSOutput(Path name) throws IOException { |
---|
271 | if (fs.exists(name)) fs.delete(name, true); |
---|
272 | FSDataOutputStream fout = fs.create(name); |
---|
273 | return fout; |
---|
274 | } |
---|
275 | |
---|
276 | /** |
---|
277 | * test none codecs |
---|
278 | */ |
---|
279 | void basicWithSomeCodec(String codec) throws IOException { |
---|
280 | Path ncTFile = new Path(ROOT, "basic.tfile"); |
---|
281 | FSDataOutputStream fout = createFSOutput(ncTFile); |
---|
282 | Writer writer = new Writer(fout, minBlockSize, codec, "memcmp", conf); |
---|
283 | writeRecords(writer); |
---|
284 | fout.close(); |
---|
285 | FSDataInputStream fin = fs.open(ncTFile); |
---|
286 | Reader reader = |
---|
287 | new Reader(fs.open(ncTFile), fs.getFileStatus(ncTFile).getLen(), conf); |
---|
288 | |
---|
289 | Scanner scanner = reader.createScanner(); |
---|
290 | readAllRecords(scanner); |
---|
291 | scanner.seekTo(getSomeKey(50)); |
---|
292 | assertTrue("location lookup failed", scanner.seekTo(getSomeKey(50))); |
---|
293 | // read the key and see if it matches |
---|
294 | byte[] readKey = readKey(scanner); |
---|
295 | assertTrue("seeked key does not match", Arrays.equals(getSomeKey(50), |
---|
296 | readKey)); |
---|
297 | |
---|
298 | scanner.seekTo(new byte[0]); |
---|
299 | byte[] val1 = readValue(scanner); |
---|
300 | scanner.seekTo(new byte[0]); |
---|
301 | byte[] val2 = readValue(scanner); |
---|
302 | assertTrue(Arrays.equals(val1, val2)); |
---|
303 | |
---|
304 | // check for lowerBound |
---|
305 | scanner.lowerBound(getSomeKey(50)); |
---|
306 | assertTrue("locaton lookup failed", scanner.currentLocation |
---|
307 | .compareTo(reader.end()) < 0); |
---|
308 | readKey = readKey(scanner); |
---|
309 | assertTrue("seeked key does not match", Arrays.equals(readKey, |
---|
310 | getSomeKey(50))); |
---|
311 | |
---|
312 | // check for upper bound |
---|
313 | scanner.upperBound(getSomeKey(50)); |
---|
314 | assertTrue("location lookup failed", scanner.currentLocation |
---|
315 | .compareTo(reader.end()) < 0); |
---|
316 | readKey = readKey(scanner); |
---|
317 | assertTrue("seeked key does not match", Arrays.equals(readKey, |
---|
318 | getSomeKey(51))); |
---|
319 | |
---|
320 | scanner.close(); |
---|
321 | // test for a range of scanner |
---|
322 | scanner = reader.createScanner(getSomeKey(10), getSomeKey(60)); |
---|
323 | readAndCheckbytes(scanner, 10, 50); |
---|
324 | assertFalse(scanner.advance()); |
---|
325 | scanner.close(); |
---|
326 | reader.close(); |
---|
327 | fin.close(); |
---|
328 | fs.delete(ncTFile, true); |
---|
329 | } |
---|
330 | |
---|
331 | // unsorted with some codec |
---|
332 | void unsortedWithSomeCodec(String codec) throws IOException { |
---|
333 | Path uTfile = new Path(ROOT, "unsorted.tfile"); |
---|
334 | FSDataOutputStream fout = createFSOutput(uTfile); |
---|
335 | Writer writer = new Writer(fout, minBlockSize, codec, null, conf); |
---|
336 | writeRecords(writer); |
---|
337 | writer.close(); |
---|
338 | fout.close(); |
---|
339 | FSDataInputStream fin = fs.open(uTfile); |
---|
340 | Reader reader = |
---|
341 | new Reader(fs.open(uTfile), fs.getFileStatus(uTfile).getLen(), conf); |
---|
342 | |
---|
343 | Scanner scanner = reader.createScanner(); |
---|
344 | readAllRecords(scanner); |
---|
345 | scanner.close(); |
---|
346 | reader.close(); |
---|
347 | fin.close(); |
---|
348 | fs.delete(uTfile, true); |
---|
349 | } |
---|
350 | |
---|
351 | public void testTFileFeatures() throws IOException { |
---|
352 | basicWithSomeCodec("none"); |
---|
353 | basicWithSomeCodec("gz"); |
---|
354 | } |
---|
355 | |
---|
356 | // test unsorted t files. |
---|
357 | public void testUnsortedTFileFeatures() throws IOException { |
---|
358 | unsortedWithSomeCodec("none"); |
---|
359 | unsortedWithSomeCodec("gz"); |
---|
360 | } |
---|
361 | |
---|
362 | private void writeNumMetablocks(Writer writer, String compression, int n) |
---|
363 | throws IOException { |
---|
364 | for (int i = 0; i < n; i++) { |
---|
365 | DataOutputStream dout = |
---|
366 | writer.prepareMetaBlock("TfileMeta" + i, compression); |
---|
367 | byte[] b = ("something to test" + i).getBytes(); |
---|
368 | dout.write(b); |
---|
369 | dout.close(); |
---|
370 | } |
---|
371 | } |
---|
372 | |
---|
373 | private void someTestingWithMetaBlock(Writer writer, String compression) |
---|
374 | throws IOException { |
---|
375 | DataOutputStream dout = null; |
---|
376 | writeNumMetablocks(writer, compression, 10); |
---|
377 | try { |
---|
378 | dout = writer.prepareMetaBlock("TfileMeta1", compression); |
---|
379 | assertTrue(false); |
---|
380 | } |
---|
381 | catch (MetaBlockAlreadyExists me) { |
---|
382 | // avoid this exception |
---|
383 | } |
---|
384 | dout = writer.prepareMetaBlock("TFileMeta100", compression); |
---|
385 | dout.close(); |
---|
386 | } |
---|
387 | |
---|
388 | private void readNumMetablocks(Reader reader, int n) throws IOException { |
---|
389 | int len = ("something to test" + 0).getBytes().length; |
---|
390 | for (int i = 0; i < n; i++) { |
---|
391 | DataInputStream din = reader.getMetaBlock("TfileMeta" + i); |
---|
392 | byte b[] = new byte[len]; |
---|
393 | din.readFully(b); |
---|
394 | assertTrue("faield to match metadata", Arrays.equals( |
---|
395 | ("something to test" + i).getBytes(), b)); |
---|
396 | din.close(); |
---|
397 | } |
---|
398 | } |
---|
399 | |
---|
400 | private void someReadingWithMetaBlock(Reader reader) throws IOException { |
---|
401 | DataInputStream din = null; |
---|
402 | readNumMetablocks(reader, 10); |
---|
403 | try { |
---|
404 | din = reader.getMetaBlock("NO ONE"); |
---|
405 | assertTrue(false); |
---|
406 | } |
---|
407 | catch (MetaBlockDoesNotExist me) { |
---|
408 | // should catch |
---|
409 | } |
---|
410 | din = reader.getMetaBlock("TFileMeta100"); |
---|
411 | int read = din.read(); |
---|
412 | assertTrue("check for status", (read == -1)); |
---|
413 | din.close(); |
---|
414 | } |
---|
415 | |
---|
416 | // test meta blocks for tfiles |
---|
417 | public void testMetaBlocks() throws IOException { |
---|
418 | Path mFile = new Path(ROOT, "meta.tfile"); |
---|
419 | FSDataOutputStream fout = createFSOutput(mFile); |
---|
420 | Writer writer = new Writer(fout, minBlockSize, "none", null, conf); |
---|
421 | someTestingWithMetaBlock(writer, "none"); |
---|
422 | writer.close(); |
---|
423 | fout.close(); |
---|
424 | FSDataInputStream fin = fs.open(mFile); |
---|
425 | Reader reader = new Reader(fin, fs.getFileStatus(mFile).getLen(), conf); |
---|
426 | someReadingWithMetaBlock(reader); |
---|
427 | fs.delete(mFile, true); |
---|
428 | reader.close(); |
---|
429 | fin.close(); |
---|
430 | } |
---|
431 | } |
---|