feat(2meet-data-optimizer): store post_status and post_date copies in post entity tables

- TMDO_Post_Status_Sync: post group tables get post_status / post_date columns (posts' collation) and an (post_status, post_date) index. Brand-new empty tables get them on creation; existing tables only via `wp tmdo post-status-sync [--dry-run] [--batch]`, never as DDL on a populated table in a web request. Per-group state in tmdo_post_status_columns ('columns' | 'ready'). - transition_post_status refreshes the copies; Hook Bus upserts that create a post row fill them in (existing rows keep theirs). - Bench (1M rows): add columns 0.05 s, index 5.3 s, batched backfill 32 s, 0 mismatches afterwards. - Verify: core unit 659 / integration 466 green, phpcs clean - PR: PR-028
2026-09-26 01:25:20 +08:00
parent f4eee2338b
commit aeca12b41d
9 changed files with 471 additions and 4 deletions
@@ -109,6 +109,9 @@ class TMDO_Core {
 		// v2.0.0: boot the unified Hook Bus when explicitly enabled.
 		TMDO_Hook_Bus_Bridge::maybe_init_hook_bus();

+		// Keep post_status / post_date copies on post entity tables current.
+		TMDO_Post_Status_Sync::register();
+
 		// Register audit log listeners for wpdo_after_write / wpdo_after_delete.
 		if ( class_exists( 'TMDO_Audit_Logger' ) ) {
 			TMDO_Audit_Logger::init();
@@ -328,15 +328,20 @@ final class TMDO_Hook_Bus {

 		try {
 			// 單一 upsert，避免 SELECT→UPDATE/INSERT 的 read-modify-write race。
+			// post：新列一併帶入 post_status / post_date 副本（只在 insert 時寫入）。
+			$extra  = 'post' === $type ? TMDO_Post_Status_Sync::insert_columns( $entity_id, $group ) : array();
 			$result = TMDO_DB::upsert(
 				$table,
-				array(
-					$id_col => $entity_id,
-					$col    => $db_value,
+				array_merge(
+					array(
+						$id_col => $entity_id,
+						$col    => $db_value,
+					),
+					$extra
 				),
 				array( $col ),
 				$id_col,
-				array( '%d', $format )
+				array_merge( array( '%d', $format ), array_fill( 0, count( $extra ), '%s' ) )
 			);

 			// 清除快取
@@ -0,0 +1,236 @@
+<?php
+/**
+ * Copies of posts.post_status / post_date on post entity tables (O-1).
+ *
+ * With the copies (and their `(post_status, post_date)` index) a routed search
+ * can filter and sort inside the flat table instead of joining back to posts.
+ *
+ * Columns only reach a table two ways, never as DDL on a populated table in a
+ * web request:
+ *   - a brand-new (empty) group table gets them when it is created;
+ *   - `wp tmdo post-status-sync` adds them to existing tables and backfills.
+ *
+ * Per-group state lives in the TMDO_Post_Status_Sync::OPTION option:
+ * 'columns' (added, backfill pending) or 'ready' (every row copied).
+ * Afterwards transition_post_status keeps the copies current, and flat writes
+ * that create a row fill them in.
+ *
+ * @package TMDO
+ */
+
+if ( ! defined( 'ABSPATH' ) ) {
+	exit;
+}
+
+/**
+ * Keeps post_status / post_date copies on post entity tables.
+ */
+final class TMDO_Post_Status_Sync {
+
+	public const OPTION = 'tmdo_post_status_columns';
+
+	private const INDEX = 'idx_post_status_date';
+
+	/**
+	 * Hook the sync into WordPress.
+	 *
+	 * @return void
+	 */
+	public static function register(): void {
+		add_action( 'transition_post_status', array( __CLASS__, 'on_transition' ), 10, 3 );
+		add_action( 'wpdo_schema_updated', array( __CLASS__, 'on_schema_updated' ), 10, 3 );
+	}
+
+	/**
+	 * '' (no columns), 'columns' or 'ready'.
+	 *
+	 * @param string $group Post group name.
+	 * @return string
+	 */
+	public static function state( string $group ): string {
+		$states = get_option( self::OPTION, array() );
+		return is_array( $states ) ? (string) ( $states[ $group ] ?? '' ) : '';
+	}
+
+	/**
+	 * Every post status change or save: refresh the copies in each table that has them.
+	 *
+	 * @param string $new_status New status.
+	 * @param string $old_status Old status.
+	 * @param object $post       Saved post.
+	 * @return void
+	 */
+	public static function on_transition( $new_status, $old_status, $post ): void {
+		global $wpdb;
+		unset( $new_status, $old_status );
+		foreach ( self::groups_with_columns() as $group ) {
+			$table = TMDO_Schema_Manager::get_table_name( 'post', $group );
+			$wpdb->query(
+				$wpdb->prepare(
+					"UPDATE `{$table}` SET `post_status` = %s, `post_date` = %s WHERE `post_id` = %d", // phpcs:ignore WordPress.DB.PreparedSQL.InterpolatedNotPrepared -- Registry table name.
+					$post->post_status,
+					$post->post_date,
+					$post->ID
+				)
+			);
+		}
+	}
+
+	/**
+	 * A post group table was just created or altered: an empty one gets the
+	 * columns now; one holding rows waits for the CLI.
+	 *
+	 * @param string $entity_type Entity type.
+	 * @param string $group       Group name.
+	 * @param string $table       Table name.
+	 * @return void
+	 */
+	public static function on_schema_updated( $entity_type, $group, $table ): void {
+		global $wpdb;
+		if ( 'post' !== $entity_type || '' !== self::state( $group ) ) {
+			return;
+		}
+		if ( null !== $wpdb->get_var( "SELECT 1 FROM `{$table}` LIMIT 1" ) ) { // phpcs:ignore WordPress.DB.PreparedSQL.InterpolatedNotPrepared -- Registry table name.
+			return;
+		}
+		if ( self::add_columns( $group ) ) {
+			self::set_state( $group, 'ready' );
+		}
+	}
+
+	/**
+	 * Copies for a row a flat write is about to create (empty when the
+	 * group's table has no such columns).
+	 *
+	 * @param int    $post_id Post ID.
+	 * @param string $group   Group name.
+	 * @return array{post_status?:string, post_date?:string}
+	 */
+	public static function insert_columns( int $post_id, string $group ): array {
+		if ( '' === self::state( $group ) ) {
+			return array();
+		}
+		$post = get_post( $post_id );
+		if ( ! $post ) {
+			return array();
+		}
+		return array(
+			'post_status' => $post->post_status,
+			'post_date'   => $post->post_date,
+		);
+	}
+
+	/**
+	 * Add the columns to every post group table and copy posts' values in.
+	 *
+	 * @param bool $dry_run Count mismatching rows only.
+	 * @param int  $batch   Rows per UPDATE.
+	 * @return array<string, array{rows:int, mismatched:int}>
+	 */
+	public static function run( bool $dry_run = false, int $batch = 1000 ): array {
+		global $wpdb;
+		$batch  = max( 1, $batch );
+		$report = array();
+		foreach ( TMDO_Entity_Registry::get_groups_for_type( 'post' ) as $group ) {
+			$table = TMDO_Schema_Manager::get_table_name( 'post', $group );
+			if ( ! TMDO_Schema_Manager::table_exists( $table ) ) {
+				continue;
+			}
+			$rows             = (int) $wpdb->get_var( "SELECT COUNT(*) FROM `{$table}`" ); // phpcs:ignore WordPress.DB.PreparedSQL.InterpolatedNotPrepared -- Registry table name.
+			$report[ $group ] = array(
+				'rows'       => $rows,
+				'mismatched' => '' === self::state( $group ) ? $rows : self::mismatches( $group ),
+			);
+			if ( $dry_run || ! self::add_columns( $group ) ) {
+				continue;
+			}
+			$max = (int) $wpdb->get_var( "SELECT MAX(id) FROM `{$table}`" ); // phpcs:ignore WordPress.DB.PreparedSQL.InterpolatedNotPrepared -- Registry table name.
+			for ( $from = 0; $from < $max; $from += $batch ) {
+				$wpdb->query(
+					$wpdb->prepare(
+						"UPDATE `{$table}` t JOIN `{$wpdb->posts}` p ON p.ID = t.post_id SET t.post_status = p.post_status, t.post_date = p.post_date WHERE t.id > %d AND t.id <= %d", // phpcs:ignore WordPress.DB.PreparedSQL.InterpolatedNotPrepared -- Registry table name.
+						$from,
+						$from + $batch
+					)
+				);
+			}
+			self::set_state( $group, 'ready' );
+		}
+		return $report;
+	}
+
+	/**
+	 * Rows whose copies differ from posts (rows of deleted posts excluded).
+	 *
+	 * @param string $group Group name.
+	 * @return int
+	 */
+	public static function mismatches( string $group ): int {
+		global $wpdb;
+		$table = TMDO_Schema_Manager::get_table_name( 'post', $group );
+		return (int) $wpdb->get_var(
+			"SELECT COUNT(*) FROM `{$table}` t JOIN `{$wpdb->posts}` p ON p.ID = t.post_id WHERE NOT ( t.post_status <=> p.post_status ) OR NOT ( t.post_date <=> p.post_date )" // phpcs:ignore WordPress.DB.PreparedSQL.InterpolatedNotPrepared -- Registry table name.
+		);
+	}
+
+	/**
+	 * Add whichever of the two columns and the index the table lacks.
+	 *
+	 * @param string $group Group name.
+	 * @return bool False when the table is missing or the ALTER failed.
+	 */
+	private static function add_columns( string $group ): bool {
+		global $wpdb;
+		$table = TMDO_Schema_Manager::get_table_name( 'post', $group );
+		if ( ! TMDO_Schema_Manager::table_exists( $table ) ) {
+			return false;
+		}
+		$columns = (array) $wpdb->get_col( "SHOW COLUMNS FROM `{$table}`" ); // phpcs:ignore WordPress.DB.PreparedSQL.InterpolatedNotPrepared -- Registry table name.
+		$indexes = array_column( (array) $wpdb->get_results( "SHOW INDEX FROM `{$table}`", ARRAY_A ), 'Key_name' ); // phpcs:ignore WordPress.DB.PreparedSQL.InterpolatedNotPrepared -- Registry table name.
+		$changes = array();
+		if ( ! in_array( 'post_status', $columns, true ) ) {
+			// Same collation as posts.post_status: the table default may differ,
+			// and comparing the two columns would then fail (illegal mix of collations).
+			$source    = $wpdb->get_row( "SHOW FULL COLUMNS FROM `{$wpdb->posts}` LIKE 'post_status'", ARRAY_A );
+			$collation = preg_replace( '/[^A-Za-z0-9_]/', '', (string) ( $source['Collation'] ?? '' ) );
+			$changes[] = 'ADD COLUMN `post_status` VARCHAR(20)' . ( '' !== $collation ? " COLLATE {$collation}" : '' ) . ' DEFAULT NULL';
+		}
+		if ( ! in_array( 'post_date', $columns, true ) ) {
+			$changes[] = 'ADD COLUMN `post_date` DATETIME DEFAULT NULL';
+		}
+		if ( ! in_array( self::INDEX, $indexes, true ) ) {
+			$changes[] = 'ADD INDEX `' . self::INDEX . '` (`post_status`, `post_date`)';
+		}
+		if ( $changes && false === $wpdb->query( "ALTER TABLE `{$table}` " . implode( ', ', $changes ) ) ) { // phpcs:ignore WordPress.DB.PreparedSQL.InterpolatedNotPrepared -- Fixed DDL on a registry table name.
+			return false;
+		}
+		if ( '' === self::state( $group ) ) {
+			self::set_state( $group, 'columns' );
+		}
+		return true;
+	}
+
+	/**
+	 * Post groups whose table has the columns.
+	 *
+	 * @return string[]
+	 */
+	private static function groups_with_columns(): array {
+		$states = get_option( self::OPTION, array() );
+		return is_array( $states ) ? array_keys( $states ) : array();
+	}
+
+	/**
+	 * Record a group's state.
+	 *
+	 * @param string $group Group name.
+	 * @param string $state 'columns' | 'ready'.
+	 * @return void
+	 */
+	private static function set_state( string $group, string $state ): void {
+		$states           = get_option( self::OPTION, array() );
+		$states           = is_array( $states ) ? $states : array();
+		$states[ $group ] = $state;
+		update_option( self::OPTION, $states, true );
+	}
+}