| 1 | <?php |
| 2 | /************************************************************* |
| 3 | # 076 License |
| 4 | |
| 5 | Copyright (c) テクニカル諏訪子 |
| 6 | |
| 7 | Permission is hereby granted to any person obtaining a copy of the software |
| 8 | Little Beast (the "Software") to use, modify, merge, copy, publish, distribute, |
| 9 | sublicense, and/or sell copies of the Software, subject to the following conditions: |
| 10 | |
| 11 | 1. **Origin Attribution**: |
| 12 | - You must not misrepresent the origin of the Software; you must not claim |
| 13 | you created the original Software. |
| 14 | - If the Software is used in a product, you must either: |
| 15 | a. Provide clear attribution in the product's documentation, user interface, |
| 16 | or other visible areas, **OR** |
| 17 | b. Pay the original developers a fee they specify in writing. |
| 18 | 2. **Usage Restriction**: |
| 19 | - The Software, or any derivative works, dependencies, or libraries |
| 20 | incorporating it, must not be used for censorship or to suppress freedom of |
| 21 | speech, expression, or creativity. Prohibited uses include, but are not |
| 22 | limited to: |
| 23 | - Censorship of so-called "hate speech", visuals, non-mainstream opinions, |
| 24 | ideas, or objective reality. |
| 25 | - Tools or systems designed to restrict access to information or |
| 26 | artistic works. |
| 27 | 3. **Notice Preservation**: |
| 28 | - This license and the above copyright notice must remain intact in all copies |
| 29 | of the source code. |
| 30 | |
| 31 | THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, |
| 32 | INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A |
| 33 | PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT |
| 34 | HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF |
| 35 | CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE |
| 36 | OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. |
| 37 | */ |
| 38 | namespace Std\Lib\Db\Sqlite; |
| 39 | |
| 40 | use Std\Lib\Db\DbException; |
| 41 | |
| 42 | /** |
| 43 | * SQLiteのBツリー(テーブル:rowidがキー、インデックス:レコードがキー) |
| 44 | * |
| 45 | * ページはPageのオブジェクトとして持ち、flush()でPagerに書く。セルは毎回ページの最後に詰めて書き直すので、 |
| 46 | * フリーブロックは使わない。ページが溢れたり少なくなったりしたら、SQLiteと同じく最大3つの兄弟のページで分け直す。 |
| 47 | */ |
| 48 | final class Btree { |
| 49 | public const TABLE_LEAF = 13; |
| 50 | public const TABLE_INTERIOR = 5; |
| 51 | public const INDEX_LEAF = 10; |
| 52 | public const INDEX_INTERIOR = 2; |
| 53 | |
| 54 | /** @var array<int, Page> */ |
| 55 | private array $pages = []; |
| 56 | private array $dirty = []; |
| 57 | private int $generation = -1; |
| 58 | /** スナップショット(文・セーブポイント):ページ番号 => [変更前のページ, 変更されていたか]、又はnull(無かった) */ |
| 59 | private array $snaps = []; |
| 60 | |
| 61 | public function __construct(private Pager $pager) {} |
| 62 | |
| 63 | /////////////// |
| 64 | // ページ |
| 65 | |
| 66 | /** ページを読む(オブジェクトのキャッシュ) */ |
| 67 | public function page(int $pgno): Page { |
| 68 | $this->sync(); |
| 69 | if ($this->snaps) $this->remember($pgno); |
| 70 | if (isset($this->pages[$pgno])) return $this->pages[$pgno]; |
| 71 | if (count($this->pages) > 3000 && empty($this->dirty)) $this->pages = []; |
| 72 | return $this->pages[$pgno] = $this->parse($pgno); |
| 73 | } |
| 74 | |
| 75 | /** Pagerのキャッシュが捨てられたら、こちらも捨てる */ |
| 76 | private function sync(): void { |
| 77 | if ($this->generation !== $this->pager->generation) { |
| 78 | $this->pages = []; |
| 79 | $this->dirty = []; |
| 80 | $this->snaps = []; |
| 81 | $this->generation = $this->pager->generation; |
| 82 | } |
| 83 | } |
| 84 | |
| 85 | /** キャッシュを捨てる(ロールバックの後) */ |
| 86 | public function reset(): void { |
| 87 | $this->pages = []; |
| 88 | $this->dirty = []; |
| 89 | $this->snaps = []; |
| 90 | } |
| 91 | |
| 92 | /** ページの変更前の状態を覚える(今のスナップショットで初めての場合) */ |
| 93 | private function remember(int $pgno): void { |
| 94 | $k = array_key_last($this->snaps); |
| 95 | if (array_key_exists($pgno, $this->snaps[$k])) return; |
| 96 | $this->snaps[$k][$pgno] = isset($this->pages[$pgno]) ? [clone $this->pages[$pgno], isset($this->dirty[$pgno])] : null; |
| 97 | } |
| 98 | |
| 99 | /** |
| 100 | * スナップショット(Pagerと一緒に使う。変更されたページをPagerに書かなくても、文だけを元に戻せる) |
| 101 | * |
| 102 | * @return int |
| 103 | */ |
| 104 | public function snapPush(): int { |
| 105 | $this->sync(); |
| 106 | $this->snaps[] = []; |
| 107 | return count($this->snaps) - 1; |
| 108 | } |
| 109 | |
| 110 | /** スナップショットを捨てる(変更は残す) */ |
| 111 | public function snapPop(?int $n): void { |
| 112 | if ($n === null) return; |
| 113 | while (count($this->snaps) > $n) { |
| 114 | $top = array_pop($this->snaps); |
| 115 | if ($this->snaps) { |
| 116 | $k = array_key_last($this->snaps); |
| 117 | $this->snaps[$k] += $top; |
| 118 | } |
| 119 | } |
| 120 | } |
| 121 | |
| 122 | /** スナップショットの時に戻す(スナップショットは残す) */ |
| 123 | public function snapRestore(?int $n): void { |
| 124 | if ($n === null || !isset($this->snaps[$n])) return; |
| 125 | $all = []; |
| 126 | for ($k = $n; $k < count($this->snaps); $k++) $all += $this->snaps[$k]; |
| 127 | foreach ($all as $pgno => $e) { |
| 128 | if ($e === null) { |
| 129 | unset($this->pages[$pgno], $this->dirty[$pgno]); |
| 130 | } else { |
| 131 | $this->pages[$pgno] = $e[0]; |
| 132 | if ($e[1]) $this->dirty[$pgno] = true; |
| 133 | else unset($this->dirty[$pgno]); |
| 134 | } |
| 135 | } |
| 136 | $this->snaps = array_slice($this->snaps, 0, $n); |
| 137 | $this->snaps[] = []; |
| 138 | } |
| 139 | |
| 140 | private function parse(int $pgno): Page { |
| 141 | if ($pgno === 1 && $this->pager->nPages === 0) { |
| 142 | // 空のファイル:sqlite_schemaは空 |
| 143 | $p = new Page(1); |
| 144 | $p->type = self::TABLE_LEAF; |
| 145 | return $p; |
| 146 | } |
| 147 | $raw = $this->pager->read($pgno); |
| 148 | $h = $pgno === 1 ? 100 : 0; |
| 149 | $p = new Page($pgno); |
| 150 | $p->type = ord($raw[$h]); |
| 151 | if (!in_array($p->type, [2, 5, 10, 13], true)) throw new DbException("データベースが壊れています(ページ{$pgno}の種類:{$p->type})", 11, 'HY000'); |
| 152 | $n = unpack('n', $raw, $h + 3)[1]; |
| 153 | $leaf = $p->isLeaf(); |
| 154 | $table = $p->isTable(); |
| 155 | if (!$leaf) $p->right = unpack('N', $raw, $h + 8)[1]; |
| 156 | $base = $h + ($leaf ? 8 : 12); |
| 157 | $ptrs = $n > 0 ? array_values(unpack("n{$n}", $raw, $base)) : []; |
| 158 | $size = 0; |
| 159 | foreach ($ptrs as $off) { |
| 160 | $pos = $off; |
| 161 | if (!$leaf) { |
| 162 | $p->left[] = unpack('N', $raw, $off)[1]; |
| 163 | $pos += 4; |
| 164 | } |
| 165 | $start = $pos; |
| 166 | if ($table && !$leaf) { |
| 167 | $p->keys[] = Record::getVarint($raw, $pos); |
| 168 | } else { |
| 169 | $len = Record::getVarint($raw, $pos); |
| 170 | if ($table) $p->keys[] = Record::getVarint($raw, $pos); |
| 171 | $local = $this->localSize($len, $table); |
| 172 | $pos += $local + ($len > $local ? 4 : 0); |
| 173 | } |
| 174 | $p->cells[] = substr($raw, $start, $pos - $start); |
| 175 | $size += $pos - $off + 2; |
| 176 | } |
| 177 | $p->size = $size; |
| 178 | if (!$table) $p->dec = $n > 0 ? array_fill(0, $n, null) : []; |
| 179 | return $p; |
| 180 | } |
| 181 | |
| 182 | private function serialize(Page $p): string { |
| 183 | $h = $p->pgno === 1 ? 100 : 0; |
| 184 | $leaf = $p->isLeaf(); |
| 185 | $off = $this->pager->usable; |
| 186 | $ptrs = ''; |
| 187 | $parts = []; |
| 188 | foreach ($p->cells as $i => $c) { |
| 189 | $cell = $leaf ? $c : pack('N', $p->left[$i]).$c; |
| 190 | $off -= strlen($cell); |
| 191 | $ptrs .= pack('n', $off); |
| 192 | $parts[] = $cell; |
| 193 | } |
| 194 | $n = count($p->cells); |
| 195 | $head = chr($p->type)."\0\0".pack('n', $n).pack('n', $off === 65536 ? 0 : $off)."\0".($leaf ? '' : pack('N', $p->right)); |
| 196 | $gap = $off - ($h + strlen($head) + strlen($ptrs)); |
| 197 | if ($gap < 0) throw new \LogicException("SQLite:ページ{$p->pgno}が溢れました"); |
| 198 | $prefix = $h ? $this->pager->header() : ''; |
| 199 | return $prefix.$head.$ptrs.str_repeat("\0", $gap).implode('', array_reverse($parts)) |
| 200 | .str_repeat("\0", $this->pager->pageSize - $this->pager->usable); |
| 201 | } |
| 202 | |
| 203 | /** 変更されたページ */ |
| 204 | private function touch(Page $p): void { |
| 205 | $this->dirty[$p->pgno] = true; |
| 206 | } |
| 207 | |
| 208 | /** 変更されたページが多ければPagerに書く(操作の間だけ。途中のページは溢れているかもしれない) */ |
| 209 | private function spill(): void { |
| 210 | if (count($this->dirty) > 500) $this->flush(); |
| 211 | } |
| 212 | |
| 213 | /** 変更されたページをPagerに書く */ |
| 214 | public function flush(): void { |
| 215 | foreach (array_keys($this->dirty) as $pgno) { |
| 216 | if (isset($this->pages[$pgno])) $this->pager->write($pgno, $this->serialize($this->pages[$pgno])); |
| 217 | } |
| 218 | $this->dirty = []; |
| 219 | } |
| 220 | |
| 221 | private function newPage(int $type): Page { |
| 222 | $this->sync(); |
| 223 | $pgno = $this->pager->allocate(); |
| 224 | if ($this->snaps) $this->remember($pgno); |
| 225 | $p = new Page($pgno); |
| 226 | $p->type = $type; |
| 227 | $this->pages[$pgno] = $p; |
| 228 | $this->touch($p); |
| 229 | return $p; |
| 230 | } |
| 231 | |
| 232 | private function freePage(int $pgno): void { |
| 233 | if ($this->snaps) $this->remember($pgno); |
| 234 | unset($this->pages[$pgno], $this->dirty[$pgno]); |
| 235 | $this->pager->free($pgno); |
| 236 | } |
| 237 | |
| 238 | /** ページに入るセルの大きさ */ |
| 239 | private function capacity(Page $p): int { |
| 240 | return $this->pager->usable - ($p->pgno === 1 ? 100 : 0) - ($p->isLeaf() ? 8 : 12); |
| 241 | } |
| 242 | |
| 243 | private function cellSize(Page $p, string $cell): int { |
| 244 | return strlen($cell) + ($p->isLeaf() ? 2 : 6); |
| 245 | } |
| 246 | |
| 247 | /////////////// |
| 248 | // ペイロードとオーバーフロー |
| 249 | |
| 250 | /** ページに置くペイロードのバイト数(残りはオーバーフローのページ) */ |
| 251 | public function localSize(int $len, bool $table): int { |
| 252 | $u = $this->pager->usable; |
| 253 | $x = $table ? $u - 35 : intdiv(($u - 12) * 64, 255) - 23; |
| 254 | if ($len <= $x) return $len; |
| 255 | $m = intdiv(($u - 12) * 32, 255) - 23; |
| 256 | $k = $m + (($len - $m) % ($u - 4)); |
| 257 | return $k <= $x ? $k : $m; |
| 258 | } |
| 259 | |
| 260 | /** セルのペイロードの全部 */ |
| 261 | public function payload(Page $p, int $i): string { |
| 262 | $c = $p->cells[$i]; |
| 263 | $pos = 0; |
| 264 | $len = Record::getVarint($c, $pos); |
| 265 | $table = $p->isTable(); |
| 266 | if ($table) Record::getVarint($c, $pos); |
| 267 | $local = $this->localSize($len, $table); |
| 268 | $data = substr($c, $pos, $local); |
| 269 | if ($len > $local) { |
| 270 | $next = unpack('N', $c, $pos + $local)[1]; |
| 271 | $rest = $len - $local; |
| 272 | $chunk = $this->pager->usable - 4; |
| 273 | while ($rest > 0) { |
| 274 | if ($next < 1) throw new DbException('データベースが壊れています(オーバーフロー)', 11, 'HY000'); |
| 275 | $o = $this->pager->read($next); |
| 276 | $take = min($rest, $chunk); |
| 277 | $data .= substr($o, 4, $take); |
| 278 | $rest -= $take; |
| 279 | $next = unpack('N', $o)[1]; |
| 280 | } |
| 281 | } |
| 282 | return $data; |
| 283 | } |
| 284 | |
| 285 | /** インデックスのセルのキー(デコードした物) */ |
| 286 | public function key(Page $p, int $i): array { |
| 287 | return $p->dec[$i] ??= Record::decode($this->payload($p, $i)); |
| 288 | } |
| 289 | |
| 290 | private function makeCell(string $payload, ?int $rowid): string { |
| 291 | $len = strlen($payload); |
| 292 | $table = $rowid !== null; |
| 293 | $head = Record::putVarint($len).($table ? Record::putVarint($rowid) : ''); |
| 294 | $local = $this->localSize($len, $table); |
| 295 | if ($local >= $len) return $head.$payload; |
| 296 | $chunk = $this->pager->usable - 4; |
| 297 | $pieces = str_split(substr($payload, $local), $chunk); |
| 298 | $pgnos = []; |
| 299 | foreach ($pieces as $_) $pgnos[] = $this->pager->allocate(); |
| 300 | foreach ($pieces as $k => $piece) { |
| 301 | $next = $pgnos[$k + 1] ?? 0; |
| 302 | $this->pager->write($pgnos[$k], str_pad(pack('N', $next).$piece, $this->pager->pageSize, "\0")); |
| 303 | } |
| 304 | return $head.substr($payload, 0, $local).pack('N', $pgnos[0]); |
| 305 | } |
| 306 | |
| 307 | /** セルのオーバーフローのページを解放する */ |
| 308 | private function freeOverflow(Page $p, string $c): void { |
| 309 | if (!$p->isLeaf() && $p->isTable()) return; |
| 310 | $pos = 0; |
| 311 | $len = Record::getVarint($c, $pos); |
| 312 | if ($p->isTable()) Record::getVarint($c, $pos); |
| 313 | $local = $this->localSize($len, $p->isTable()); |
| 314 | if ($len <= $local) return; |
| 315 | $next = unpack('N', $c, $pos + $local)[1]; |
| 316 | $n = intdiv($len - $local + $this->pager->usable - 5, $this->pager->usable - 4); |
| 317 | for ($k = 0; $k < $n && $next > 0; $k++) { |
| 318 | $o = $this->pager->read($next); |
| 319 | $this->freePage($next); |
| 320 | $next = unpack('N', $o)[1]; |
| 321 | } |
| 322 | } |
| 323 | |
| 324 | /////////////// |
| 325 | // テーブル |
| 326 | |
| 327 | /** 新しいテーブル(又はインデックス)のツリー */ |
| 328 | public function create(bool $table): int { |
| 329 | return $this->newPage($table ? self::TABLE_LEAF : self::INDEX_LEAF)->pgno; |
| 330 | } |
| 331 | |
| 332 | private static function lowerBound(array $keys, int $k): int { |
| 333 | $lo = 0; |
| 334 | $hi = count($keys); |
| 335 | while ($lo < $hi) { |
| 336 | $mid = ($lo + $hi) >> 1; |
| 337 | if ($keys[$mid] < $k) $lo = $mid + 1; |
| 338 | else $hi = $mid; |
| 339 | } |
| 340 | return $lo; |
| 341 | } |
| 342 | |
| 343 | private function child(Page $p, int $i): int { |
| 344 | return $i < count($p->cells) ? $p->left[$i] : $p->right; |
| 345 | } |
| 346 | |
| 347 | /** @return array [path, 葉, 位置, 有るか] */ |
| 348 | private function seek(int $root, int $rowid): array { |
| 349 | $path = []; |
| 350 | $p = $this->page($root); |
| 351 | while (!$p->isLeaf()) { |
| 352 | $i = self::lowerBound($p->keys, $rowid); |
| 353 | $path[] = [$p, $i]; |
| 354 | $p = $this->page($this->child($p, $i)); |
| 355 | } |
| 356 | $i = self::lowerBound($p->keys, $rowid); |
| 357 | return [$path, $p, $i, $i < count($p->keys) && $p->keys[$i] === $rowid]; |
| 358 | } |
| 359 | |
| 360 | /** |
| 361 | * 行を読む |
| 362 | * |
| 363 | * @return string|null レコード(無ければnull) |
| 364 | */ |
| 365 | public function get(int $root, int $rowid): ?string { |
| 366 | [, $p, $i, $found] = $this->seek($root, $rowid); |
| 367 | return $found ? $this->payload($p, $i) : null; |
| 368 | } |
| 369 | |
| 370 | /** 行を入れる(同じrowidが有れば置き換える) */ |
| 371 | public function insert(int $root, int $rowid, string $record): void { |
| 372 | $this->spill(); |
| 373 | [$path, $p, $i, $found] = $this->seek($root, $rowid); |
| 374 | $cell = $this->makeCell($record, $rowid); |
| 375 | if ($found) { |
| 376 | $this->freeOverflow($p, $p->cells[$i]); |
| 377 | $p->size += strlen($cell) - strlen($p->cells[$i]); |
| 378 | $p->cells[$i] = $cell; |
| 379 | } else { |
| 380 | array_splice($p->cells, $i, 0, [$cell]); |
| 381 | array_splice($p->keys, $i, 0, [$rowid]); |
| 382 | $p->size += strlen($cell) + 2; |
| 383 | } |
| 384 | $this->touch($p); |
| 385 | $this->fix($path, $p); |
| 386 | } |
| 387 | |
| 388 | /** 行を消す */ |
| 389 | public function delete(int $root, int $rowid): bool { |
| 390 | $this->spill(); |
| 391 | [$path, $p, $i, $found] = $this->seek($root, $rowid); |
| 392 | if (!$found) return false; |
| 393 | $this->freeOverflow($p, $p->cells[$i]); |
| 394 | $p->size -= strlen($p->cells[$i]) + 2; |
| 395 | array_splice($p->cells, $i, 1); |
| 396 | array_splice($p->keys, $i, 1); |
| 397 | $this->touch($p); |
| 398 | $this->fix($path, $p); |
| 399 | return true; |
| 400 | } |
| 401 | |
| 402 | /** 一番大きいrowid(空なら0) */ |
| 403 | public function maxRowid(int $root): int { |
| 404 | $p = $this->page($root); |
| 405 | while (!$p->isLeaf()) $p = $this->page($p->right); |
| 406 | return empty($p->keys) ? 0 : end($p->keys); |
| 407 | } |
| 408 | |
| 409 | /** |
| 410 | * テーブルの行を順番に |
| 411 | * |
| 412 | * @param int $root |
| 413 | * @param int|null $from 最小のrowid |
| 414 | * @param int|null $to 最大のrowid |
| 415 | * @return \Generator rowid => レコード |
| 416 | */ |
| 417 | public function scan(int $root, ?int $from = null, ?int $to = null): \Generator { |
| 418 | $stop = false; |
| 419 | yield from $this->walkTable($root, $from, $to, $stop); |
| 420 | } |
| 421 | |
| 422 | private function walkTable(int $pgno, ?int $from, ?int $to, bool &$stop): \Generator { |
| 423 | $p = $this->page($pgno); |
| 424 | $n = count($p->cells); |
| 425 | $i = $from === null ? 0 : self::lowerBound($p->keys, $from); |
| 426 | if ($p->isLeaf()) { |
| 427 | for (; $i < $n; $i++) { |
| 428 | if ($to !== null && $p->keys[$i] > $to) { |
| 429 | $stop = true; |
| 430 | return; |
| 431 | } |
| 432 | yield $p->keys[$i] => $this->payload($p, $i); |
| 433 | } |
| 434 | return; |
| 435 | } |
| 436 | for (; $i <= $n; $i++) { |
| 437 | yield from $this->walkTable($this->child($p, $i), $from, $to, $stop); |
| 438 | if ($stop || ($to !== null && $i < $n && $p->keys[$i] >= $to)) { |
| 439 | $stop = true; |
| 440 | return; |
| 441 | } |
| 442 | } |
| 443 | } |
| 444 | |
| 445 | /** 行の数 */ |
| 446 | public function count(int $root): int { |
| 447 | $p = $this->page($root); |
| 448 | if ($p->isLeaf()) return count($p->cells); |
| 449 | $n = 0; |
| 450 | for ($i = 0; $i <= count($p->cells); $i++) $n += $this->count($this->child($p, $i)); |
| 451 | return $n; |
| 452 | } |
| 453 | |
| 454 | /** |
| 455 | * ツリーを空にする($dropならルートも解放する) |
| 456 | */ |
| 457 | public function clear(int $root, bool $drop = false): void { |
| 458 | $p = $this->page($root); |
| 459 | $this->freeSubtree($p, false); |
| 460 | if ($drop) { |
| 461 | $this->freePage($root); |
| 462 | return; |
| 463 | } |
| 464 | $p->type = $p->isTable() ? self::TABLE_LEAF : self::INDEX_LEAF; |
| 465 | $p->cells = $p->left = $p->keys = $p->dec = []; |
| 466 | $p->right = 0; |
| 467 | $p->size = 0; |
| 468 | $this->touch($p); |
| 469 | } |
| 470 | |
| 471 | private function freeSubtree(Page $p, bool $self): void { |
| 472 | foreach ($p->cells as $c) $this->freeOverflow($p, $c); |
| 473 | if (!$p->isLeaf()) { |
| 474 | for ($i = 0; $i <= count($p->cells); $i++) $this->freeSubtree($this->page($this->child($p, $i)), true); |
| 475 | } |
| 476 | if ($self) $this->freePage($p->pgno); |
| 477 | } |
| 478 | |
| 479 | /////////////// |
| 480 | // インデックス |
| 481 | |
| 482 | /** |
| 483 | * 最初の位置(キーの前方一致で比べる) |
| 484 | * |
| 485 | * @param bool $strict trueなら「>」、falseなら「≧」 |
| 486 | */ |
| 487 | private function bound(Page $p, array $key, array $idx, bool $strict): int { |
| 488 | $lo = 0; |
| 489 | $hi = count($p->cells); |
| 490 | while ($lo < $hi) { |
| 491 | $mid = ($lo + $hi) >> 1; |
| 492 | $c = Record::compareKeys($this->key($p, $mid), $key, $idx['colls'], $idx['desc']); |
| 493 | if ($c < 0 || ($strict && $c === 0)) $lo = $mid + 1; |
| 494 | else $hi = $mid; |
| 495 | } |
| 496 | return $lo; |
| 497 | } |
| 498 | |
| 499 | /** |
| 500 | * インデックスにキーを入れる |
| 501 | * |
| 502 | * @param array $idx root、colls、desc(rowidの列も含む) |
| 503 | * @param array $key 列の値…、rowid |
| 504 | */ |
| 505 | public function indexInsert(array $idx, array $key): void { |
| 506 | $this->spill(); |
| 507 | $path = []; |
| 508 | $p = $this->page($idx['root']); |
| 509 | while (true) { |
| 510 | $i = $this->bound($p, $key, $idx, false); |
| 511 | if ($i < count($p->cells) && count($this->key($p, $i)) === count($key) |
| 512 | && Record::compareKeys($this->key($p, $i), $key, $idx['colls'], $idx['desc']) === 0) return; |
| 513 | if ($p->isLeaf()) break; |
| 514 | $path[] = [$p, $i]; |
| 515 | $p = $this->page($this->child($p, $i)); |
| 516 | } |
| 517 | $cell = $this->makeCell(Record::encode($key), null); |
| 518 | array_splice($p->cells, $i, 0, [$cell]); |
| 519 | array_splice($p->dec, $i, 0, [$key]); |
| 520 | $p->size += strlen($cell) + 2; |
| 521 | $this->touch($p); |
| 522 | $this->fix($path, $p); |
| 523 | } |
| 524 | |
| 525 | /** |
| 526 | * キーの前方一致の項目を探す(一意の制約の確認用) |
| 527 | * |
| 528 | * @param int|null $self 除くrowid |
| 529 | * @return int|null 見つけた項目のrowid |
| 530 | */ |
| 531 | public function indexFind(array $idx, array $prefix, ?int $self = null): ?int { |
| 532 | $p = $this->page($idx['root']); |
| 533 | while (true) { |
| 534 | $i = $this->bound($p, $prefix, $idx, false); |
| 535 | if ($i < count($p->cells)) { |
| 536 | $k = $this->key($p, $i); |
| 537 | if (Record::compareKeys($k, $prefix, $idx['colls'], $idx['desc']) === 0) { |
| 538 | $id = end($k); |
| 539 | if ($id !== $self) return $id; |
| 540 | // 自分自身:他にも有るかもしれないので全部を見る |
| 541 | foreach ($this->indexScan($idx, $prefix, true, $prefix, true) as $k2) { |
| 542 | if (end($k2) !== $self) return end($k2); |
| 543 | } |
| 544 | return null; |
| 545 | } |
| 546 | } |
| 547 | if ($p->isLeaf()) return null; |
| 548 | $p = $this->page($this->child($p, $i)); |
| 549 | } |
| 550 | } |
| 551 | |
| 552 | /** インデックスからキーを消す */ |
| 553 | public function indexDelete(array $idx, array $key): bool { |
| 554 | $this->spill(); |
| 555 | $path = []; |
| 556 | $p = $this->page($idx['root']); |
| 557 | while (true) { |
| 558 | $i = $this->bound($p, $key, $idx, false); |
| 559 | if ($i < count($p->cells) && Record::compareKeys($this->key($p, $i), $key, $idx['colls'], $idx['desc']) === 0) break; |
| 560 | if ($p->isLeaf()) return false; |
| 561 | $path[] = [$p, $i]; |
| 562 | $p = $this->page($this->child($p, $i)); |
| 563 | } |
| 564 | $this->decAll($p); |
| 565 | if ($p->isLeaf()) { |
| 566 | $this->freeOverflow($p, $p->cells[$i]); |
| 567 | $p->size -= strlen($p->cells[$i]) + 2; |
| 568 | array_splice($p->cells, $i, 1); |
| 569 | array_splice($p->dec, $i, 1); |
| 570 | $this->touch($p); |
| 571 | $this->fix($path, $p); |
| 572 | return true; |
| 573 | } |
| 574 | |
| 575 | // 内部のページ:左のツリーの一番大きいキーと置き換える |
| 576 | $pathL = array_merge($path, [[$p, $i]]); |
| 577 | $q = $this->page($p->left[$i]); |
| 578 | while (!$q->isLeaf()) { |
| 579 | $pathL[] = [$q, count($q->cells)]; |
| 580 | $q = $this->page($q->right); |
| 581 | } |
| 582 | $this->decAll($q); |
| 583 | $last = count($q->cells) - 1; |
| 584 | $cell = $q->cells[$last]; |
| 585 | $dec = $this->key($q, $last); |
| 586 | $q->size -= strlen($cell) + 2; |
| 587 | array_splice($q->cells, $last, 1); |
| 588 | array_splice($q->dec, $last, 1); |
| 589 | $this->freeOverflow($p, $p->cells[$i]); |
| 590 | $p->size += strlen($cell) - strlen($p->cells[$i]); |
| 591 | $p->cells[$i] = $cell; |
| 592 | $p->dec[$i] = $dec; |
| 593 | $this->touch($p); |
| 594 | $this->touch($q); |
| 595 | if ($this->fix($pathL, $q) > count($path)) $this->fix($path, $p); |
| 596 | return true; |
| 597 | } |
| 598 | |
| 599 | /** decの配列をセルと同じ長さにする(array_spliceの前に) */ |
| 600 | private function decAll(Page $p): void { |
| 601 | for ($i = 0, $n = count($p->cells); $i < $n; $i++) $p->dec[$i] ??= null; |
| 602 | ksort($p->dec); |
| 603 | } |
| 604 | |
| 605 | /** |
| 606 | * インデックスの範囲 |
| 607 | * |
| 608 | * @param array $idx |
| 609 | * @param array|null $lo 最小のキー(前方一致) |
| 610 | * @param bool $loInc 最小を含む |
| 611 | * @param array|null $hi 最大のキー(前方一致) |
| 612 | * @param bool $hiInc 最大を含む |
| 613 | * @return \Generator キー(列の値…、rowid) |
| 614 | */ |
| 615 | public function indexScan(array $idx, ?array $lo = null, bool $loInc = true, ?array $hi = null, bool $hiInc = true): \Generator { |
| 616 | $stop = false; |
| 617 | yield from $this->walkIndex($idx['root'], $idx, $lo, $loInc, $hi, $hiInc, $stop); |
| 618 | } |
| 619 | |
| 620 | private function walkIndex(int $pgno, array $idx, ?array $lo, bool $loInc, ?array $hi, bool $hiInc, bool &$stop): \Generator { |
| 621 | $p = $this->page($pgno); |
| 622 | $n = count($p->cells); |
| 623 | $i = $lo === null ? 0 : $this->bound($p, $lo, $idx, !$loInc); |
| 624 | $leaf = $p->isLeaf(); |
| 625 | for (; $i <= $n; $i++) { |
| 626 | if (!$leaf) { |
| 627 | yield from $this->walkIndex($this->child($p, $i), $idx, $lo, $loInc, $hi, $hiInc, $stop); |
| 628 | if ($stop) return; |
| 629 | } |
| 630 | if ($i === $n) return; |
| 631 | $k = $this->key($p, $i); |
| 632 | if ($hi !== null) { |
| 633 | $c = Record::compareKeys($k, $hi, $idx['colls'], $idx['desc']); |
| 634 | if ($c > 0 || (!$hiInc && $c === 0)) { |
| 635 | $stop = true; |
| 636 | return; |
| 637 | } |
| 638 | } |
| 639 | yield $k; |
| 640 | } |
| 641 | } |
| 642 | |
| 643 | /////////////// |
| 644 | // 釣り合い |
| 645 | |
| 646 | /** |
| 647 | * 変更したページを直す(溢れた・少なくなったら兄弟と分け直す) |
| 648 | * |
| 649 | * @param array $path 親のページと位置の配列(ルートから) |
| 650 | * @param Page $p 変更したページ |
| 651 | * @return int 確認を終えた深さ(パスの長さ) |
| 652 | */ |
| 653 | private function fix(array $path, Page $p): int { |
| 654 | while (true) { |
| 655 | $cap = $this->capacity($p); |
| 656 | $over = $p->size > $cap; |
| 657 | if (empty($path)) { |
| 658 | if ($over) { |
| 659 | $c = $this->deeper($p); |
| 660 | $path = [[$p, 0]]; |
| 661 | $p = $c; |
| 662 | continue; |
| 663 | } |
| 664 | if (!$p->isLeaf() && empty($p->cells)) $this->shallower($p); |
| 665 | return 0; |
| 666 | } |
| 667 | $under = $p->size < intdiv($cap, 3); |
| 668 | if (!$over && !$under) return count($path); |
| 669 | [$parent, $idx] = array_pop($path); |
| 670 | if (!$over && empty($parent->cells)) { |
| 671 | // 兄弟が無い |
| 672 | $p = $parent; |
| 673 | continue; |
| 674 | } |
| 675 | $this->balance($parent, $idx); |
| 676 | $p = $parent; |
| 677 | } |
| 678 | } |
| 679 | |
| 680 | /** ルートが溢れた:中身を新しいページに移して、ルートを1段上にする */ |
| 681 | private function deeper(Page $root): Page { |
| 682 | $c = $this->newPage($root->type); |
| 683 | $c->cells = $root->cells; |
| 684 | $c->left = $root->left; |
| 685 | $c->keys = $root->keys; |
| 686 | $c->dec = $root->dec; |
| 687 | $c->right = $root->right; |
| 688 | $c->size = $root->size; |
| 689 | $root->type = $root->isTable() ? self::TABLE_INTERIOR : self::INDEX_INTERIOR; |
| 690 | $root->cells = $root->left = $root->keys = $root->dec = []; |
| 691 | $root->size = 0; |
| 692 | $root->right = $c->pgno; |
| 693 | $this->touch($root); |
| 694 | return $c; |
| 695 | } |
| 696 | |
| 697 | /** ルートの子が1つだけ:子の中身をルートに入れる(入れば) */ |
| 698 | private function shallower(Page $root): void { |
| 699 | $c = $this->page($root->right); |
| 700 | $cap = $this->pager->usable - ($root->pgno === 1 ? 100 : 0) - ($c->isLeaf() ? 8 : 12); |
| 701 | if ($c->size > $cap) return; |
| 702 | $this->decAll($c); |
| 703 | $root->type = $c->type; |
| 704 | $root->cells = $c->cells; |
| 705 | $root->left = $c->left; |
| 706 | $root->keys = $c->keys; |
| 707 | $root->dec = $c->dec; |
| 708 | $root->right = $c->right; |
| 709 | $root->size = $c->size; |
| 710 | $this->freePage($c->pgno); |
| 711 | $this->touch($root); |
| 712 | } |
| 713 | |
| 714 | /** 親の$idx番目の子とその兄弟(最大3つ)のセルを分け直す */ |
| 715 | private function balance(Page $parent, int $idx): void { |
| 716 | $n = count($parent->cells); |
| 717 | $first = max(0, $idx - 1); |
| 718 | $last = min($n, $first + 2); |
| 719 | $first = max(0, $last - 2); |
| 720 | $sibs = []; |
| 721 | for ($j = $first; $j <= $last; $j++) $sibs[] = $this->page($this->child($parent, $j)); |
| 722 | $leaf = $sibs[0]->isLeaf(); |
| 723 | $table = $sibs[0]->isTable(); |
| 724 | $type = $sibs[0]->type; |
| 725 | if (!$table) $this->decAll($parent); |
| 726 | |
| 727 | // 全部のセル(間の区切りも) |
| 728 | $items = []; |
| 729 | $lastRight = 0; |
| 730 | $ns = count($sibs); |
| 731 | foreach ($sibs as $j => $s) { |
| 732 | if ($s->isLeaf() !== $leaf) throw new DbException('データベースが壊れています(ツリーの深さ)', 11, 'HY000'); |
| 733 | foreach ($s->cells as $i => $c) { |
| 734 | $items[] = [$c, $leaf ? 0 : $s->left[$i], $table ? $s->keys[$i] : 0, $table ? null : ($s->dec[$i] ?? null)]; |
| 735 | } |
| 736 | if (!$leaf) $lastRight = $s->right; |
| 737 | if ($j < $ns - 1) { |
| 738 | $pi = $first + $j; |
| 739 | if ($table && $leaf) continue; |
| 740 | if ($table) { |
| 741 | $items[] = [Record::putVarint($parent->keys[$pi]), $s->right, $parent->keys[$pi], null]; |
| 742 | } else { |
| 743 | $items[] = [$parent->cells[$pi], $leaf ? 0 : $s->right, 0, $parent->dec[$pi] ?? null]; |
| 744 | } |
| 745 | } |
| 746 | } |
| 747 | |
| 748 | $cap = $this->pager->usable - ($leaf ? 8 : 12); |
| 749 | $extra = $leaf ? 2 : 6; |
| 750 | [$groups, $divs] = $this->distribute($items, $cap, !($table && $leaf), $extra); |
| 751 | $k = count($groups); |
| 752 | |
| 753 | $pgs = []; |
| 754 | for ($j = 0; $j < $k; $j++) $pgs[] = $j < $ns ? $sibs[$j] : $this->newPage($type); |
| 755 | for ($j = $k; $j < $ns; $j++) $this->freePage($sibs[$j]->pgno); |
| 756 | |
| 757 | foreach ($groups as $j => $g) { |
| 758 | $pg = $pgs[$j]; |
| 759 | $pg->type = $type; |
| 760 | $pg->cells = $pg->left = $pg->keys = $pg->dec = []; |
| 761 | $pg->size = 0; |
| 762 | foreach ($g as $it) { |
| 763 | $pg->cells[] = $it[0]; |
| 764 | if (!$leaf) $pg->left[] = $it[1]; |
| 765 | if ($table) $pg->keys[] = $it[2]; |
| 766 | else $pg->dec[] = $it[3]; |
| 767 | $pg->size += strlen($it[0]) + $extra; |
| 768 | } |
| 769 | if (!$leaf) $pg->right = $j < $k - 1 ? $divs[$j][1] : $lastRight; |
| 770 | $this->touch($pg); |
| 771 | } |
| 772 | |
| 773 | // 親の区切り |
| 774 | $cells = $left = $keys = $dec = []; |
| 775 | for ($j = 0; $j < $k - 1; $j++) { |
| 776 | if ($table) { |
| 777 | $key = $leaf ? end($groups[$j])[2] : $divs[$j][2]; |
| 778 | $cells[] = Record::putVarint($key); |
| 779 | $keys[] = $key; |
| 780 | } else { |
| 781 | $cells[] = $divs[$j][0]; |
| 782 | $dec[] = $divs[$j][3]; |
| 783 | } |
| 784 | $left[] = $pgs[$j]->pgno; |
| 785 | } |
| 786 | array_splice($parent->cells, $first, $ns - 1, $cells); |
| 787 | array_splice($parent->left, $first, $ns - 1, $left); |
| 788 | if ($table) array_splice($parent->keys, $first, $ns - 1, $keys); |
| 789 | else array_splice($parent->dec, $first, $ns - 1, $dec); |
| 790 | $slot = $first + $k - 1; |
| 791 | if ($slot < count($parent->cells)) $parent->left[$slot] = $pgs[$k - 1]->pgno; |
| 792 | else $parent->right = $pgs[$k - 1]->pgno; |
| 793 | $parent->size = 0; |
| 794 | foreach ($parent->cells as $c) $parent->size += strlen($c) + 6; |
| 795 | $this->touch($parent); |
| 796 | } |
| 797 | |
| 798 | /** |
| 799 | * セルをページに分ける |
| 800 | * |
| 801 | * @param bool $promote グループの間のセルを親の区切りにする(テーブルの葉以外) |
| 802 | * @return array [グループの配列, 区切りの配列] |
| 803 | */ |
| 804 | private function distribute(array $items, int $cap, bool $promote, int $extra): array { |
| 805 | $sizes = []; |
| 806 | $total = 0; |
| 807 | $max = 0; |
| 808 | foreach ($items as $it) { |
| 809 | $s = strlen($it[0]) + $extra; |
| 810 | $sizes[] = $s; |
| 811 | $total += $s; |
| 812 | $max = max($max, $s); |
| 813 | } |
| 814 | $pack = function(int $limit) use ($items, $sizes, $promote): array { |
| 815 | $groups = $divs = $cur = []; |
| 816 | $cs = 0; |
| 817 | foreach ($items as $i => $it) { |
| 818 | if ($cur && $cs + $sizes[$i] > $limit) { |
| 819 | $groups[] = $cur; |
| 820 | $cur = []; |
| 821 | $cs = 0; |
| 822 | if ($promote) { |
| 823 | $divs[] = $it; |
| 824 | continue; |
| 825 | } |
| 826 | } |
| 827 | $cur[] = $it; |
| 828 | $cs += $sizes[$i]; |
| 829 | } |
| 830 | if ($cur || empty($groups)) { |
| 831 | $groups[] = $cur; |
| 832 | } else if ($promote) { |
| 833 | $groups[] = [array_pop($divs)]; |
| 834 | } |
| 835 | return [$groups, $divs]; |
| 836 | }; |
| 837 | $res = $pack($cap); |
| 838 | $k = count($res[0]); |
| 839 | if ($k > 1) { |
| 840 | // なるべく同じ大きさに |
| 841 | $target = max($max, intdiv($total, $k)); |
| 842 | $step = max(1, intdiv($cap - $target, 8)); |
| 843 | for ($limit = $target; $limit < $cap; $limit += $step) { |
| 844 | $try = $pack($limit); |
| 845 | if (count($try[0]) === $k) { |
| 846 | $res = $try; |
| 847 | break; |
| 848 | } |
| 849 | } |
| 850 | } |
| 851 | return $res; |
| 852 | } |
| 853 | } |