就是对着规范里面的每一个小节进行编码实现 。先看二进制布局 ， 由于使用128 bit去存储 ， 可以借鉴UUID那样 ， 使用两个long类似的成员变量存储ULID的信息 ， 看起来像这样：
public final class ULID {/** The most significant 64 bits of this ULID.**/private final long msb;/** The least significant 64 bits of this ULID.**/private final long lsb;public ULID(long msb, long lsb) {this.msb = msb;this.lsb = lsb;}}按照ULID的组成来看 ， 可以提供一个入参为时间戳和随机数字节数组的构造：
public ULID(long timestamp, byte[] randomness) {if ((timestamp & TIMESTAMP_MASK) != 0) {throw new IllegalArgumentException("Invalid timestamp");}if (Objects.isNull(randomness) || RANDOMNESS_BYTE_LEN != randomness.length) {throw new IllegalArgumentException("Invalid randomness");}long msb = 0;long lsb = 0;// 时间戳左移16位 ， 低位补零准备填入部分随机数位 ， 即16_bit_uint_randommsb |= timestamp << 16;// randomness[0]左移0位填充到16_bit_uint_random的高8位 ， randomness[1]填充到16_bit_uint_random的低8位msb |= (long) (randomness[0x0] & 0xff) << 8;// randomness[1]填充到16_bit_uint_random的低8位msb |= randomness[0x1] & 0xff;// randomness[2] ~ randomness[9]填充到剩余的bit_uint_random中 ， 要左移相应的位lsb |= (long) (randomness[0x2] & 0xff) << 56;lsb |= (long) (randomness[0x3] & 0xff) << 48;lsb |= (long) (randomness[0x4] & 0xff) << 40;lsb |= (long) (randomness[0x5] & 0xff) << 32;lsb |= (long) (randomness[0x6] & 0xff) << 24;lsb |= (long) (randomness[0x7] & 0xff) << 16;lsb |= (long) (randomness[0x8] & 0xff) << 8;lsb |= (randomness[0x9] & 0xff);this.msb = msb;this.lsb = lsb;}这是完全按照规范的二进制布局编写代码 ， 可以像UUID的构造那样精简一下：
long msb = 0;long lsb = 0;byte[] data = https://www.huyubaike.com/biancheng/new byte[16];byte[] ts = ByteBuffer.allocate(8).putLong(0, timestamp << 16).array();System.arraycopy(ts, 0, data, 0, 6);System.arraycopy(randomness, 0, data, 6, 10);for (int i = 0; i < 8; i++)msb = (msb << 8) | (data[i] & 0xff);for (int i = 8; i < 16; i++)lsb = (lsb << 8) | (data[i] & 0xff);接着可以简单添加下面几个方法：
public long getMostSignificantBits() {return this.msb;}public long getLeastSignificantBits() {return this.lsb;}// 静态工厂方法 ， 由UUID实例生成ULID实例public static ULID fromUUID(UUID uuid) {return new ULID(uuid.getMostSignificantBits(), uuid.getLeastSignificantBits());}// 实例方法 ， 当前ULID实例转换为UUID实例public UUID toUUID() {return new UUID(this.msb, this.lsb);}接着需要覆盖toString()方法 ， 这是ULID的核心方法 ， 需要通过Crockford Base32编码生成规范的字符串表示形式 。由于128 bit要映射为26 char ， 这里可以考虑分三段进行映射 ， 也就是48 bit时间戳映射为10 char ， 剩下的两部分随机数分别做40 bit到8 char的映射 ， 加起来就是26 char：
|----------||----------------|TimestampRandomness[split to 2 part]48bit => 10char80bit => 16char编写方法：
/*** Default alphabet of ULID*/private static final char[] DEFAULT_ALPHABET = {'0', '1', '2', '3', '4', '5', '6', '7', '8', '9', 'A', 'B', 'C','D', 'E', 'F', 'G', 'H', 'J', 'K', 'M', 'N', 'P', 'Q', 'R', 'S', 'T', 'V', 'W', 'X', 'Y', 'Z'};/*** Default alphabet mask*/private static final int DEFAULT_ALPHABET_MASK = 0b11111;/*** Character num of ULID*/private static final int ULID_CHAR_LEN = 0x1a;@Overridepublic String toString() {return toCanonicalString(DEFAULT_ALPHABET);}public String toCanonicalString(char[] alphabet) {char[] chars = new char[ULID_CHAR_LEN];long timestamp = this.msb >> 16;// 第一部分随机数取msb的低16位+lsb的高24位 ， 这里(msb & 0xffff) << 24作为第一部分随机数的高16位 ， 所以要左移24位long randMost = ((this.msb & 0xffffL) << 24) | (this.lsb >>> 40);// 第二部分随机数取lsb的低40位 ， 0xffffffffffL是2^40-1long randLeast = (this.lsb & 0xffffffffffL);// 接着每个部分的偏移量和DEFAULT_ALPHABET_MASK(31)做一次或运算就行 ， 就是char[index] = alphabet[(part >> (step * index)) & 31]chars[0x00] = alphabet[(int) (timestamp >>> 45 & DEFAULT_ALPHABET_MASK)];chars[0x01] = alphabet[(int) (timestamp >>> 40 & DEFAULT_ALPHABET_MASK)];chars[0x02] = alphabet[(int) (timestamp >>> 35 & DEFAULT_ALPHABET_MASK)];chars[0x03] = alphabet[(int) (timestamp >>> 30 & DEFAULT_ALPHABET_MASK)];chars[0x04] = alphabet[(int) (timestamp >>> 25 & DEFAULT_ALPHABET_MASK)];chars[0x05] = alphabet[(int) (timestamp >>> 20 & DEFAULT_ALPHABET_MASK)];chars[0x06] = alphabet[(int) (timestamp >>> 15 & DEFAULT_ALPHABET_MASK)];chars[0x07] = alphabet[(int) (timestamp >>> 10 & DEFAULT_ALPHABET_MASK)];chars[0x08] = alphabet[(int) (timestamp >>> 5 & DEFAULT_ALPHABET_MASK)];chars[0x09] = alphabet[(int) (timestamp & DEFAULT_ALPHABET_MASK)];chars[0x0a] = alphabet[(int) (randMost >>> 35 & DEFAULT_ALPHABET_MASK)];chars[0x0b] = alphabet[(int) (randMost >>> 30 & DEFAULT_ALPHABET_MASK)];chars[0x0c] = alphabet[(int) (randMost >>> 25 & DEFAULT_ALPHABET_MASK)];chars[0x0d] = alphabet[(int) (randMost >>> 20 & DEFAULT_ALPHABET_MASK)];chars[0x0e] = alphabet[(int) (randMost >>> 15 & DEFAULT_ALPHABET_MASK)];chars[0x0f] = alphabet[(int) (randMost >>> 10 & DEFAULT_ALPHABET_MASK)];chars[0x10] = alphabet[(int) (randMost >>> 5 & DEFAULT_ALPHABET_MASK)];chars[0x11] = alphabet[(int) (randMost & DEFAULT_ALPHABET_MASK)];chars[0x12] = alphabet[(int) (randLeast >>> 35 & DEFAULT_ALPHABET_MASK)];chars[0x13] = alphabet[(int) (randLeast >>> 30 & DEFAULT_ALPHABET_MASK)];chars[0x14] = alphabet[(int) (randLeast >>> 25 & DEFAULT_ALPHABET_MASK)];chars[0x15] = alphabet[(int) (randLeast >>> 20 & DEFAULT_ALPHABET_MASK)];chars[0x16] = alphabet[(int) (randLeast >>> 15 & DEFAULT_ALPHABET_MASK)];chars[0x17] = alphabet[(int) (randLeast >>> 10 & DEFAULT_ALPHABET_MASK)];chars[0x18] = alphabet[(int) (randLeast >>> 5 & DEFAULT_ALPHABET_MASK)];chars[0x19] = alphabet[(int) (randLeast & DEFAULT_ALPHABET_MASK)];return new String(chars);}