Hadoop源码分析-Text
Text是Hadoop中的一个Writable类,定义了Hadoop中的其中的数据类型以及操作。
This class stores text using standard UTF8 encoding. It provides methods to serialize, deserialize, and compare texts at byte level. The type of length is integer and is serialized using zero-compressed format.
In addition, it provides methods for string traversal without converting the byte array to a string.Also includes utilities for serializing/deserialing a string,encoding / decoding a string, checking if a byte array contains valid UTF8 code, calculating the length of an encoded string.
由上图的Writable层次结构图可以看到绝大多数的数据类型都实现了Writable、WritableComparable接口,在此先分析一下这两个接口情况。自顶下下逐步分析。
Writable接口的定义如下:
1 package org.apache.hadoop.io; 2 3 import java.io.DataOutput; 4 import java.io.DataInput; 5 import java.io.IOException; 6 public interface Writable { 7 void write(DataOutput out) throws IOException; 8 void readFields(DataInput in) throws IOException; 9 }
void write(DataOutput out) throws IOException /*object将自身字段序列化后的的字节流写入输出流out中。 参数:out - 接收object序列化后的字节流的输出流. */
void readFields(DataInput in) throws IOException /*将输入流in中的字节流反序列化然后写入object的字段 参数:字节流的出处 */
而DataInput、DataOutput是java.io.*中最基本的输入输出流接口,其他输入输出流都需要实现DataInput与DataOutput这两个接口的方法。关于这两个接口,另外开篇分析解读。
到此Writable接口解读完毕,其实这些东西大家看看API文档也可以看懂的,我只是想详细了解一下Writable类所以就写一次更加明白。
WritableComparable接口定义如下:
package org.apache.hadoop.io; public interface WritableComparable<T> extends Writable, comparable<T> { }
咋一看这个WritableComparable没有方法,其实它的方法全都是通过继承而来的,Writable接口上面已经分析了,所以WritableComparable以下两个方法。
void write(DataOutput out) throws IOException;
void readFields(DataInput in) throws IOException;
还有来自comparable的方法,comparable是属于java.lang.*中的一个接口,它只有一个方法。
int compareTo( T other); /*比较此对象与指定对象other的顺序。如果该对象小于、等于或大于指定对象,则分别返回负整数、零或正整数。参数:o - 要比较的对象。返回:负整数、零或正整数,根据此对象是小于、等于还是大于指定对象。 */
简单来说实现WritableComparable的类是一个可写可比较的类。
现在来分析基本类Text,声明定义如下
public class Text extends BinaryComparable implements WritableComparable<BinaryComparable>;
1 package org.apache.hadoop.io; 2 3 import java.io.IOException; 4 import java.io.DataInput; 5 import java.io.DataOutput; 6 import java.nio.ByteBuffer; 7 import java.nio.CharBuffer; 8 import java.nio.charset.CharacterCodingException; 9 import java.nio.charset.Charset; 10 import java.nio.charset.CharsetDecoder; 11 import java.nio.charset.CharsetEncoder; 12 import java.nio.charset.CodingErrorAction; 13 import java.nio.charset.MalformedInputException; 14 import java.text.CharacterIterator; 15 import java.text.StringCharacterIterator; 16 17 import org.apache.commons.logging.Log; 18 import org.apache.commons.logging.LogFactory; 19 20 /** This class stores text using standard UTF8 encoding. It provides methods 21 * to serialize, deserialize, and compare texts at byte level. The type of 22 * length is integer and is serialized using zero-compressed format. <p>In 23 * addition, it provides methods for string traversal without converting the 24 * byte array to a string. <p>Also includes utilities for 25 * serializing/deserialing a string, coding/decoding a string, checking if a 26 * byte array contains valid UTF8 code, calculating the length of an encoded 27 * string. 28 */ 29 public class Text extends BinaryComparable 30 implements WritableComparable<BinaryComparable> { 31 private static final Log LOG= LogFactory.getLog(Text.class); 32 33 private static ThreadLocal<CharsetEncoder> ENCODER_FACTORY = 34 new ThreadLocal<CharsetEncoder>() { 35 protected CharsetEncoder initialValue() { 36 return Charset.forName("UTF-8").newEncoder(). 37 onMalformedInput(CodingErrorAction.REPORT). 38 onUnmappableCharacter(CodingErrorAction.REPORT); 39 } 40 }; 41 42 private static ThreadLocal<CharsetDecoder> DECODER_FACTORY = 43 new ThreadLocal<CharsetDecoder>() { 44 protected CharsetDecoder initialValue() { 45 return Charset.forName("UTF-8").newDecoder(). 46 onMalformedInput(CodingErrorAction.REPORT). 47 onUnmappableCharacter(CodingErrorAction.REPORT); 48 } 49 }; 50 51 private static final byte [] EMPTY_BYTES = new byte[0]; 52 53 private byte[] bytes; 54 private int length; 55 56 public Text() { 57 bytes = EMPTY_BYTES; 58 } 59 60 /** Construct from a string. 61 */ 62 public Text(String string) { 63 set(string); 64 } 65 66 /** Construct from another text. */ 67 public Text(Text utf8) { 68 set(utf8); 69 } 70 71 /** Construct from a byte array. 72 */ 73 public Text(byte[] utf8) { 74 set(utf8); 75 } 76 77 /** 78 * Returns the raw bytes; however, only data up to {@link #getLength()} is 79 * valid. 80 */ 81 public byte[] getBytes() { 82 return bytes; 83 } 84 85 /** Returns the number of bytes in the byte array */ 86 public int getLength() { 87 return length; 88 } 89 90 /** 91 * Returns the Unicode Scalar Value (32-bit integer value) 92 * for the character at <code>position</code>. Note that this 93 * method avoids using the converter or doing String instatiation 94 * @return the Unicode scalar value at position or -1 95 * if the position is invalid or points to a 96 * trailing byte 97 */ 98 public int charAt(int position) { 99 if (position > this.length) return -1; // too long 100 if (position < 0) return -1; // duh. 101 102 ByteBuffer bb = (ByteBuffer)ByteBuffer.wrap(bytes).position(position); 103 return bytesToCodePoint(bb.slice()); 104 } 105 106 public int find(String what) { 107 return find(what, 0); 108 } 109 110 /** 111 * Finds any occurence of <code>what</code> in the backing 112 * buffer, starting as position <code>start</code>. The starting 113 * position is measured in bytes and the return value is in 114 * terms of byte position in the buffer. The backing buffer is 115 * not converted to a string for this operation. 116 * @return byte position of the first occurence of the search 117 * string in the UTF-8 buffer or -1 if not found 118 */ 119 public int find(String what, int start) { 120 try { 121 ByteBuffer src = ByteBuffer.wrap(this.bytes,0,this.length); 122 ByteBuffer tgt = encode(what); 123 byte b = tgt.get(); 124 src.position(start); 125 126 while (src.hasRemaining()) { 127 if (b == src.get()) { // matching first byte 128 src.mark(); // save position in loop 129 tgt.mark(); // save position in target 130 boolean found = true; 131 int pos = src.position()-1; 132 while (tgt.hasRemaining()) { 133 if (!src.hasRemaining()) { // src expired first 134 tgt.reset(); 135 src.reset(); 136 found = false; 137 break; 138 } 139 if (!(tgt.get() == src.get())) { 140 tgt.reset(); 141 src.reset(); 142 found = false; 143 break; // no match 144 } 145 } 146 if (found) return pos; 147 } 148 } 149 return -1; // not found 150 } catch (CharacterCodingException e) { 151 // can't get here 152 e.printStackTrace(); 153 return -1; 154 } 155 } 156 /** Set to contain the contents of a string. 157 */ 158 public void set(String string) { 159 try { 160 ByteBuffer bb = encode(string, true); 161 bytes = bb.array(); 162 length = bb.limit(); 163 }catch(CharacterCodingException e) { 164 throw new RuntimeException("Should not have happened " + e.toString()); 165 } 166 } 167 168 /** Set to a utf8 byte array 169 */ 170 public void set(byte[] utf8) { 171 set(utf8, 0, utf8.length); 172 } 173 174 /** copy a text. */ 175 public void set(Text other) { 176 set(other.getBytes(), 0, other.getLength()); 177 } 178 179 /** 180 * Set the Text to range of bytes 181 * @param utf8 the data to copy from 182 * @param start the first position of the new string 183 * @param len the number of bytes of the new string 184 */ 185 public void set(byte[] utf8, int start, int len) { 186 setCapacity(len, false); 187 System.arraycopy(utf8, start, bytes, 0, len); 188 this.length = len; 189 } 190 191 /** 192 * Append a range of bytes to the end of the given text 193 * @param utf8 the data to copy from 194 * @param start the first position to append from utf8 195 * @param len the number of bytes to append 196 */ 197 public void append(byte[] utf8, int start, int len) { 198 setCapacity(length + len, true); 199 System.arraycopy(utf8, start, bytes, length, len); 200 length += len; 201 } 202 203 /** 204 * Clear the string to empty. 205 */ 206 public void clear() { 207 length = 0; 208 } 209 210 /* 211 * Sets the capacity of this Text object to <em>at least</em> 212 * <code>len</code> bytes. If the current buffer is longer, 213 * then the capacity and existing content of the buffer are 214 * unchanged. If <code>len</code> is larger 215 * than the current capacity, the Text object's capacity is 216 * increased to match. 217 * @param len the number of bytes we need 218 * @param keepData should the old data be kept 219 */ 220 private void setCapacity(int len, boolean keepData) { 221 if (bytes == null || bytes.length < len) { 222 byte[] newBytes = new byte[len]; 223 if (bytes != null && keepData) { 224 System.arraycopy(bytes, 0, newBytes, 0, length); 225 } 226 bytes = newBytes; 227 } 228 } 229 230 /** 231 * Convert text back to string 232 * @see java.lang.Object#toString() 233 */ 234 public String toString() { 235 try { 236 return decode(bytes, 0, length); 237 } catch (CharacterCodingException e) { 238 throw new RuntimeException("Should not have happened " + e.toString()); 239 } 240 } 241 242 /** deserialize 243 */ 244 public void readFields(DataInput in) throws IOException { 245 int newLength = WritableUtils.readVInt(in); 246 setCapacity(newLength, false); 247 in.readFully(bytes, 0, newLength); 248 length = newLength; 249 } 250 251 /** Skips over one Text in the input. */ 252 public static void skip(DataInput in) throws IOException { 253 int length = WritableUtils.readVInt(in); 254 WritableUtils.skipFully(in, length); 255 } 256 257 /** serialize 258 * write this object to out 259 * length uses zero-compressed encoding 260 * @see Writable#write(DataOutput) 261 */ 262 public void write(DataOutput out) throws IOException { 263 WritableUtils.writeVInt(out, length); 264 out.write(bytes, 0, length); 265 } 266 267 /** Returns true iff <code>o</code> is a Text with the same contents. */ 268 public boolean equals(Object o) { 269 if (o instanceof Text) 270 return super.equals(o); 271 return false; 272 } 273 274 public int hashCode() { 275 return super.hashCode(); 276 } 277 278 /** A WritableComparator optimized for Text keys. */ 279 public static class Comparator extends WritableComparator { 280 public Comparator() { 281 super(Text.class); 282 } 283 284 public int compare(byte[] b1, int s1, int l1, 285 byte[] b2, int s2, int l2) { 286 int n1 = WritableUtils.decodeVIntSize(b1[s1]); 287 int n2 = WritableUtils.decodeVIntSize(b2[s2]); 288 return compareBytes(b1, s1+n1, l1-n1, b2, s2+n2, l2-n2); 289 } 290 } 291 292 static { 293 // register this comparator 294 WritableComparator.define(Text.class, new Comparator()); 295 } 296 297 /// STATIC UTILITIES FROM HERE DOWN 298 /** 299 * Converts the provided byte array to a String using the 300 * UTF-8 encoding. If the input is malformed, 301 * replace by a default value. 302 */ 303 public static String decode(byte[] utf8) throws CharacterCodingException { 304 return decode(ByteBuffer.wrap(utf8), true); 305 } 306 307 public static String decode(byte[] utf8, int start, int length) 308 throws CharacterCodingException { 309 return decode(ByteBuffer.wrap(utf8, start, length), true); 310 } 311 312 /** 313 * Converts the provided byte array to a String using the 314 * UTF-8 encoding. If <code>replace</code> is true, then 315 * malformed input is replaced with the 316 * substitution character, which is U+FFFD. Otherwise the 317 * method throws a MalformedInputException. 318 */ 319 public static String decode(byte[] utf8, int start, int length, boolean replace) 320 throws CharacterCodingException { 321 return decode(ByteBuffer.wrap(utf8, start, length), replace); 322 } 323 324 private static String decode(ByteBuffer utf8, boolean replace) 325 throws CharacterCodingException { 326 CharsetDecoder decoder = DECODER_FACTORY.get(); 327 if (replace) { 328 decoder.onMalformedInput( 329 java.nio.charset.CodingErrorAction.REPLACE); 330 decoder.onUnmappableCharacter(CodingErrorAction.REPLACE); 331 } 332 String str = decoder.decode(utf8).toString(); 333 // set decoder back to its default value: REPORT 334 if (replace) { 335 decoder.onMalformedInput(CodingErrorAction.REPORT); 336 decoder.onUnmappableCharacter(CodingErrorAction.REPORT); 337 } 338 return str; 339 } 340 341 /** 342 * Converts the provided String to bytes using the 343 * UTF-8 encoding. If the input is malformed, 344 * invalid chars are replaced by a default value. 345 * @return ByteBuffer: bytes stores at ByteBuffer.array() 346 * and length is ByteBuffer.limit() 347 */ 348 349 public static ByteBuffer encode(String string) 350 throws CharacterCodingException { 351 return encode(string, true); 352 } 353 354 /** 355 * Converts the provided String to bytes using the 356 * UTF-8 encoding. If <code>replace</code> is true, then 357 * malformed input is replaced with the 358 * substitution character, which is U+FFFD. Otherwise the 359 * method throws a MalformedInputException. 360 * @return ByteBuffer: bytes stores at ByteBuffer.array() 361 * and length is ByteBuffer.limit() 362 */ 363 public static ByteBuffer encode(String string, boolean replace) 364 throws CharacterCodingException { 365 CharsetEncoder encoder = ENCODER_FACTORY.get(); 366 if (replace) { 367 encoder.onMalformedInput(CodingErrorAction.REPLACE); 368 encoder.onUnmappableCharacter(CodingErrorAction.REPLACE); 369 } 370 ByteBuffer bytes = 371 encoder.encode(CharBuffer.wrap(string.toCharArray())); 372 if (replace) { 373 encoder.onMalformedInput(CodingErrorAction.REPORT); 374 encoder.onUnmappableCharacter(CodingErrorAction.REPORT); 375 } 376 return bytes; 377 } 378 379 /** Read a UTF8 encoded string from in 380 */ 381 public static String readString(DataInput in) throws IOException { 382 int length = WritableUtils.readVInt(in); 383 byte [] bytes = new byte[length]; 384 in.readFully(bytes, 0, length); 385 return decode(bytes); 386 } 387 388 /** Write a UTF8 encoded string to out 389 */ 390 public static int writeString(DataOutput out, String s) throws IOException { 391 ByteBuffer bytes = encode(s); 392 int length = bytes.limit(); 393 WritableUtils.writeVInt(out, length); 394 out.write(bytes.array(), 0, length); 395 return length; 396 } 397 398 // states for validateUTF8 399 400 private static final int LEAD_BYTE = 0; 401 402 private static final int TRAIL_BYTE_1 = 1; 403 404 private static final int TRAIL_BYTE = 2; 405 406 /** 407 * Check if a byte array contains valid utf-8 408 * @param utf8 byte array 409 * @throws MalformedInputException if the byte array contains invalid utf-8 410 */ 411 public static void validateUTF8(byte[] utf8) throws MalformedInputException { 412 validateUTF8(utf8, 0, utf8.length); 413 } 414 415 /** 416 * Check to see if a byte array is valid utf-8 417 * @param utf8 the array of bytes 418 * @param start the offset of the first byte in the array 419 * @param len the length of the byte sequence 420 * @throws MalformedInputException if the byte array contains invalid bytes 421 */ 422 public static void validateUTF8(byte[] utf8, int start, int len) 423 throws MalformedInputException { 424 int count = start; 425 int leadByte = 0; 426 int length = 0; 427 int state = LEAD_BYTE; 428 while (count < start+len) { 429 int aByte = ((int) utf8[count] & 0xFF); 430 431 switch (state) { 432 case LEAD_BYTE: 433 leadByte = aByte; 434 length = bytesFromUTF8[aByte]; 435 436 switch (length) { 437 case 0: // check for ASCII 438 if (leadByte > 0x7F) 439 throw new MalformedInputException(count); 440 break; 441 case 1: 442 if (leadByte < 0xC2 || leadByte > 0xDF) 443 throw new MalformedInputException(count); 444 state = TRAIL_BYTE_1; 445 break; 446 case 2: 447 if (leadByte < 0xE0 || leadByte > 0xEF) 448 throw new MalformedInputException(count); 449 state = TRAIL_BYTE_1; 450 break; 451 case 3: 452 if (leadByte < 0xF0 || leadByte > 0xF4) 453 throw new MalformedInputException(count); 454 state = TRAIL_BYTE_1; 455 break; 456 default: 457 // too long! Longest valid UTF-8 is 4 bytes (lead + three) 458 // or if < 0 we got a trail byte in the lead byte position 459 throw new MalformedInputException(count); 460 } // switch (length) 461 break; 462 463 case TRAIL_BYTE_1: 464 if (leadByte == 0xF0 && aByte < 0x90) 465 throw new MalformedInputException(count); 466 if (leadByte == 0xF4 && aByte > 0x8F) 467 throw new MalformedInputException(count); 468 if (leadByte == 0xE0 && aByte < 0xA0) 469 throw new MalformedInputException(count); 470 if (leadByte == 0xED && aByte > 0x9F) 471 throw new MalformedInputException(count); 472 // falls through to regular trail-byte test!! 473 case TRAIL_BYTE: 474 if (aByte < 0x80 || aByte > 0xBF) 475 throw new MalformedInputException(count); 476 if (--length == 0) { 477 state = LEAD_BYTE; 478 } else { 479 state = TRAIL_BYTE; 480 } 481 break; 482 } // switch (state) 483 count++; 484 } 485 } 486 487 /** 488 * Magic numbers for UTF-8. These are the number of bytes 489 * that <em>follow</em> a given lead byte. Trailing bytes 490 * have the value -1. The values 4 and 5 are presented in 491 * this table, even though valid UTF-8 cannot include the 492 * five and six byte sequences. 493 */ 494 static final int[] bytesFromUTF8 = 495 { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 496 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 497 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 498 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 499 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 500 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 501 0, 0, 0, 0, 0, 0, 0, 502 // trail bytes 503 -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 504 -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 505 -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 506 -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 1, 1, 1, 1, 1, 507 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 508 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 3, 509 3, 3, 3, 3, 3, 3, 3, 4, 4, 4, 4, 5, 5, 5, 5 }; 510 511 /** 512 * Returns the next code point at the current position in 513 * the buffer. The buffer's position will be incremented. 514 * Any mark set on this buffer will be changed by this method! 515 */ 516 public static int bytesToCodePoint(ByteBuffer bytes) { 517 bytes.mark(); 518 byte b = bytes.get(); 519 bytes.reset(); 520 int extraBytesToRead = bytesFromUTF8[(b & 0xFF)]; 521 if (extraBytesToRead < 0) return -1; // trailing byte! 522 int ch = 0; 523 524 switch (extraBytesToRead) { 525 case 5: ch += (bytes.get() & 0xFF); ch <<= 6; /* remember, illegal UTF-8 */ 526 case 4: ch += (bytes.get() & 0xFF); ch <<= 6; /* remember, illegal UTF-8 */ 527 case 3: ch += (bytes.get() & 0xFF); ch <<= 6; 528 case 2: ch += (bytes.get() & 0xFF); ch <<= 6; 529 case 1: ch += (bytes.get() & 0xFF); ch <<= 6; 530 case 0: ch += (bytes.get() & 0xFF); 531 } 532 ch -= offsetsFromUTF8[extraBytesToRead]; 533 534 return ch; 535 } 536 537 538 static final int offsetsFromUTF8[] = 539 { 0x00000000, 0x00003080, 540 0x000E2080, 0x03C82080, 0xFA082080, 0x82082080 }; 541 542 /** 543 * For the given string, returns the number of UTF-8 bytes 544 * required to encode the string. 545 * @param string text to encode 546 * @return number of UTF-8 bytes required to encode 547 */ 548 public static int utf8Length(String string) { 549 CharacterIterator iter = new StringCharacterIterator(string); 550 char ch = iter.first(); 551 int size = 0; 552 while (ch != CharacterIterator.DONE) { 553 if ((ch >= 0xD800) && (ch < 0xDC00)) { 554 // surrogate pair? 555 char trail = iter.next(); 556 if ((trail > 0xDBFF) && (trail < 0xE000)) { 557 // valid pair 558 size += 4; 559 } else { 560 // invalid pair 561 size += 3; 562 iter.previous(); // rewind one 563 } 564 } else if (ch < 0x80) { 565 size++; 566 } else if (ch < 0x800) { 567 size += 2; 568 } else { 569 // ch < 0x10000, that is, the largest char value 570 size += 3; 571 } 572 ch = iter.next(); 573 } 574 return size; 575 } 576 }
它继承了BinaryComparable基类、实现了WritableComparable<BinaryComparable>接口
WritableComparable已经在上面讲述,现来分析BinaryComparable基类,定义如下:
1 package org.apache.hadoop.io; 2 public abstract class BinaryComparable implements Comparable<BinaryComparable> { 3 public abstract int getLength(); 4 public abstract byte[] getBytes(); 5 public int compareTo(BinaryComparable other) { 6 if (this == other) 7 return 0; 8 return WritableComparator.compareBytes(getBytes(), 0, getLength(), 9 other.getBytes(), 0, other.getLength()); 10 } 11 public int compareTo(byte[] other, int off, int len) { 12 return WritableComparator.compareBytes(getBytes(), 0, getLength(), 13 other, off, len); 14 } 15 public boolean equals(Object other) { 16 if (!(other instanceof BinaryComparable)) 17 return false; 18 BinaryComparable that = (BinaryComparable)other; 19 if (this.getLength() != that.getLength()) 20 return false; 21 return this.compareTo(that) == 0; 22 } 23 public int hashCode() { 24 return WritableComparator.hashBytes(getBytes(), getLength()); 25 } 26 27 }
BinaryComparable是一个抽象类,主要是提供一个在二进制流这一层次直接比较两个对象的功能
其中
WritableComparator.compareBytes(getBytes(), 0, getLength(), other.getBytes(), 0, other.getLength());
是根据字典序排序返回比较结果。
而
WritableComparator.hashBytes(getBytes(), getLength());
则是返回字节流的hashCode;
现在总括看看Text的方法
Text是针对UTF-8序列的Writable类,一般可以认为它等价于java.lang.String 的 Writable,为了与输入流输出流DataInput、DataOutput兼容(DataInput与DataOutput是使用UTF-8修改版进行编码的),Text是使用Java的UTF-8修改版来进行编码。关于UTF-8修改版如下:
虽然Text使用UTF-8修改版编码,但是它着重还是使用了UTF-8编码,因此Text类与Java的String之间存在差异。
1、索引 对于Text类的索引是根据编码后的字节序列中的位置实现的,并非字符串中的Unicode字符,也不是Java Char的编码单元(String),对于ASCII字符串,这三个索引位置的概念是一致的,因为在三个编码方式当中,ASCII的编码大小均为一个字节。
但是Unicode使用多个字节来进行编码时,Text与String的差异就出现了。
import org.apache.hadoop.io.Text;public class TextExample {public static void main(String[] args) {// TODO Auto-generated method stubString str = new String("\u0041\u00DF\u6771\uD801\uDC00");Text t = new Text(str);System.out.println(str.length());System.out.println(t.getLength());} }
输出
5
10
证实了String的长度是其所含char编码单位的个数,但是Text的长度却是其UTF-8编码的字节数(10=1+2+3+4),这个怎么算出来呢?
待续
Hadoop源码分析-Text相关推荐
- Hadoop源码分析(12)
Hadoop源码分析(12) 1. journalnode客户端 在文档(11)中分析了初始化editlog的方法.在初始化之前其会根据集 群的配置状态选择不同的方式来进行初始化.在HA状态下,其 ...
- Hadoop源码分析(25)
Hadoop源码分析(25) ZKFC源码分析 从文档(4)到文档(24),详细分析了namenode的启动流程.最后namenode会以standby模式启动.但在standby模式下的name ...
- Hadoop源码分析(四)
2021SC@SDUSC 研究内容简略介绍 上周我们分析了org.apache.hadoop.mapreduce.Cluster中的的核心代码,本周将继续对mapreduce部分进行分析.在对Clus ...
- Hadoop源码分析-Context
学编程第一个肯定是hello world,Hadoop也不例外,它的hello world就是Wordcount,单词统计例子 1 package org.apache.hadoop.examples ...
- hadoop源码分析_Spark2.x精通:Job触发流程源码深度剖析(一)
, 一.概述 之前几篇文章对Spark集群的Master.Worker启动流程进行了源码剖析,后面直接从客户端角度出发,讲解了spark-submit任务提交过程及driver的启动:集群启动.任务 ...
- Hadoop 源码分析(二四)FSNamesystem
以下轮到FSNamesystem 出场了. FSNamesystem.java 一共同拥有4573 行.而整个namenode 文件夹下全部的Java 程序总共也仅仅有16876 行,把FSNames ...
- Hadoop源码分析笔记(十一):数据节点--数据节点整体运行
数据节点整体运行 数据节点通过数据节点存储和文件系统数据集,管理着保存在Linux文件系统上的数据块,通过流式接口提供数据块的读.写.替换.复制和校验信息等功能.建立在上述基础上的数据节点,还需要维护 ...
- Hadoop源码分析HDFS ClientProtocol——getBlockLocations
2019独角兽企业重金招聘Python工程师标准>>> Class ClientProtocol是HDFS客户端与NameNode之间的接口.Client通过调用ClientProt ...
- Hadoop源码分析16: IPC流程(11) 整体流程
1.发送请求:query.getCPUStatus("Intel"); 2.传送到:RPCInvoker.invoke(Object proxy, Method method, O ...
最新文章
- HttpServletResponse对象(一)
- Ardino基础教程 1_让Arduino说出HelloWorld
- MongoDB在Linux下常用优化设置
- iOS XIB等比例适配
- 拦截方法并替换成自己的方法
- Mysql数据类型TINYINT(1)与BOOLEAN踩坑记
- python安装失败错误代码0x_详解雨林木风Win10提示“Python 0x80070643安装时发生严重错误”解决方法...
- app_offline.htm的作用
- 第一个IOS应用终于审核通过了,好开心!
- linux内核make 时间久,Linux内核makefile问题
- 网络安全实验5 证书应用和证书管理
- php十六进制转二进制,php 十六进制转化为二进制的代码
- Problem A. Alarm Clock
- [Android6.0][MTK6737] 修改默认允许位置来源第三方应用安装
- 教学中常用的计算机软件及作用,信息技术教学中常用的教学方法有哪些
- 可长时间佩戴的耳机真的存在吗?骨传导耳机是否对耳朵伤害更小?
- wwbizsrv.exe-应用程序错误
- 手机上那些排版优雅的文章是怎么实现的?
- 解决XAMPP中无法勾选/无服务/modules无反应/Service灰框/无✓X/提供虚拟服务器
- Spark面试,Spark面试题,Spark面试汇总
热门文章
- 【Windows 逆向】使用 CE 工具挖掘关键数据内存真实地址 ( 完整流程演示 | 查找临时内存地址 | 查找真实指针地址 )
- 【错误记录】Android Studio 编译报错 ( SDK location not found )
- 【错误记录】Kotlin 编译报错 ( Not nullable value required to call an ‘iterator()‘ method on for-loop range )
- 网页分享到微信、微博、QQ空间、百度贴吧等
- 7-1 银行业务队列简单模拟 (25 分)
- lnmp环境切换php版本,并安装相应redis扩展
- Cadence 建立封装:多个引脚于芯片内部连接的封装建立方式
- UIRemoteNotificationType 参考
- 如何使用HTML5,CSS3和PHP创建一个联系表格
- [转]window.open居中