编辑文件

编辑文件：FilterData.php

<?php

declare(strict_types=1);

namespace Automattic\WooCommerce\Internal\ProductFilters;

use Automattic\WooCommerce\Internal\ProductFilters\Interfaces\QueryClausesGenerator;
use Automattic\WooCommerce\Internal\ProductFilters\TaxonomyHierarchyData;
use WC_Cache_Helper;

defined( 'ABSPATH' ) || exit;

/**
 * Class for filter counts.
 *
 * @internal For exclusive usage of WooCommerce core, backwards compatibility not guaranteed.
 */
class FilterData {
	/**
	 * Instance of QueryClauses.
	 *
	 * @var QueryClausesGenerator
	 */
	private $query_clauses;

/**
	 * Instance of TaxonomyHierarchyData.
	 *
	 * @var TaxonomyHierarchyData
	 */
	private $taxonomy_hierarchy_data;

/**
	 * Constructor.
	 *
	 * @param QueryClausesGenerator $query_clauses Instance of QueryClausesGenerator.
	 * @param TaxonomyHierarchyData $taxonomy_hierarchy_data Instance of TaxonomyHierarchyData.
	 */
	public function __construct( QueryClausesGenerator $query_clauses, TaxonomyHierarchyData $taxonomy_hierarchy_data ) {
		$this->query_clauses           = $query_clauses;
		$this->taxonomy_hierarchy_data = $taxonomy_hierarchy_data;
	}

/**
	 * Get price data for current products.
	 *
	 * @param array $query_vars The WP_Query arguments.
	 * @return object
	 */
	public function get_filtered_price( array $query_vars ) {
		/**
		 * Allows offloading the filter data to external services like Elasticsearch.
		 *
		 * @hook woocommerce_pre_product_filter_data
		 *
		 * @since 9.9.0
		 *
		 * @param array  $results      The results for current query.
		 * @param string $filter_type  The type of filter. Accepts price|stock|rating|attribute.
		 * @param array  $query_vars   The query arguments to calculate the filter data.
		 * @param array  $extra        Some filter types require extra arguments for calculation, like attribute.
		 * @return array The filtered results or null to continue with default processing.
		 */
		$pre_filter_counts = apply_filters( 'woocommerce_pre_product_filter_data', null, 'price', $query_vars, array() );

if ( is_array( $pre_filter_counts ) ) {
			return $pre_filter_counts;
		}

$transient_key = $this->get_transient_key( $query_vars, 'price' );
		$cached_data   = $this->get_cache( $transient_key );

if ( ! empty( $cached_data ) ) {
			return $cached_data;
		}

$results     = array();
		$product_ids = $this->get_cached_product_ids( $query_vars );

if ( $product_ids ) {
			global $wpdb;

$price_filter_sql = "
			SELECT min( min_price ) as min_price, MAX( max_price ) as max_price
			FROM {$wpdb->wc_product_meta_lookup}
			WHERE product_id IN ( {$product_ids} )
			";

/**
			* We can't use $wpdb->prepare() here because using %s with
			* $wpdb->prepare() for a subquery won't work as it will escape the SQL
			* query.
			* We're using the query as is, same as Core does.
			*/
			// phpcs:ignore WordPress.DB.PreparedSQL.NotPrepared
			$results = (array) $wpdb->get_row( $price_filter_sql );
		}

/**
		 * Filters the product filter data before it is returned.
		 *
		 * @hook woocommerce_product_filter_data
		 * @since 9.9.0
		 *
		 * @param array  $results      The results for current query.
		 * @param string $filter_type  The type of filter. Accepts price|stock|rating|attribute.
		 * @param array  $query_vars   The query arguments to calculate the filter data.
		 * @param array  $extra        Some filter types require extra arguments for calculation, like attribute.
		 * @return array The filtered results
		 */
		$results = apply_filters( 'woocommerce_product_filter_data', $results, 'price', $query_vars, array() );

$this->set_cache( $transient_key, $results );

return $results;
	}

/**
	 * Get stock status counts for the current products.
	 *
	 * @param array $query_vars The WP_Query arguments.
	 * @param array $statuses   Array of stock status values to count.
	 * @return array status=>count pairs.
	 */
	public function get_stock_status_counts( array $query_vars, array $statuses ) {
		/**
		 * Filter the data. @see get_filtered_price() for full documentation.
		 */
		$pre_filter_counts = apply_filters( 'woocommerce_pre_product_filter_data', null, 'stock', $query_vars, array() ); // phpcs:ignore WooCommerce.Commenting.CommentHooks.MissingSinceComment

if ( is_array( $pre_filter_counts ) ) {
			return $pre_filter_counts;
		}

$transient_key = $this->get_transient_key( $query_vars, 'stock' );
		$cached_data   = $this->get_cache( $transient_key );

if ( ! empty( $cached_data ) ) {
			return $cached_data;
		}

$results     = array();
		$product_ids = $this->get_cached_product_ids( $query_vars );

if ( $product_ids ) {
			global $wpdb;

if ( get_option( 'woocommerce_product_lookup_table_is_generating' ) ) {
				// Optimization note: this serves as a fallback while wc_product_meta_lookup is being populated and is bypassed most of the time.
				$sql = "
					SELECT meta_value AS stock_status, COUNT( DISTINCT post_id ) AS status_count
					FROM {$wpdb->postmeta}
					WHERE post_id IN ( {$product_ids} ) AND meta_key = '_stock_status'
					GROUP BY meta_value
				";
			} else {
				// Optimization note: this is the main performance driver as the database processes fewer rows than when scanning the posts meta table.
				$sql = "
					SELECT stock_status, COUNT( DISTINCT product_id ) as status_count
					FROM {$wpdb->wc_product_meta_lookup}
					WHERE product_id IN ( {$product_ids} )
					GROUP BY stock_status
				";
			}

$results = array_fill_keys( $statuses, 0 );
			foreach ( $wpdb->get_results( $sql ) as $row ) { // phpcs:ignore WordPress.DB.PreparedSQL.NotPrepared
				if ( isset( $results[ $row->stock_status ] ) ) {
					$results[ $row->stock_status ] = (int) $row->status_count;
				}
			}
		}

/**
		 * Filter the results. @see get_filtered_price() for full documentation.
		 */
		$results = apply_filters( 'woocommerce_product_filter_data', $results, 'stock', $query_vars, array() ); // phpcs:ignore WooCommerce.Commenting.CommentHooks.MissingSinceComment

$this->set_cache( $transient_key, $results );

return $results;
	}

/**
	 * Get rating counts for the current products.
	 *
	 * @param array $query_vars The WP_Query arguments.
	 * @return array rating=>count pairs.
	 */
	public function get_rating_counts( array $query_vars ) {
		/**
		 * Filter the data. @see get_filtered_price() for full documentation.
		 */
		$pre_filter_counts = apply_filters( 'woocommerce_pre_product_filter_data', null, 'rating', $query_vars, array() ); // phpcs:ignore WooCommerce.Commenting.CommentHooks.MissingSinceComment

if ( is_array( $pre_filter_counts ) ) {
			return $pre_filter_counts;
		}

$transient_key = $this->get_transient_key( $query_vars, 'rating' );
		$cached_data   = $this->get_cache( $transient_key );

if ( ! empty( $cached_data ) ) {
			return $cached_data;
		}

$results     = array();
		$product_ids = $this->get_cached_product_ids( $query_vars );

if ( $product_ids ) {
			global $wpdb;

$rating_count_sql = "
				SELECT COUNT( DISTINCT product_id ) as product_count, ROUND( average_rating, 0 ) as rounded_average_rating
				FROM {$wpdb->wc_product_meta_lookup}
				WHERE product_id IN ( {$product_ids} )
				AND average_rating > 0
				GROUP BY rounded_average_rating
				ORDER BY rounded_average_rating DESC
			";

/**
			* We can't use $wpdb->prepare() here because using %s with
			* $wpdb->prepare() for a subquery won't work as it will escape the
			* SQL query.
			* We're using the query as is, same as Core does.
			*/
			// phpcs:ignore WordPress.DB.PreparedSQL.NotPrepared
			$results = $wpdb->get_results( $rating_count_sql );
			$results = array_map( 'absint', wp_list_pluck( $results, 'product_count', 'rounded_average_rating' ) );
		}

/**
		 * Filter the results. @see get_filtered_price() for full documentation.
		 */
		$results = apply_filters( 'woocommerce_product_filter_data', $results, 'rating', $query_vars, array() ); // phpcs:ignore WooCommerce.Commenting.CommentHooks.MissingSinceComment

$this->set_cache( $transient_key, $results );

return $results;
	}

/**
	 * Get attribute counts for the current products.
	 *
	 * @param array  $query_vars         The WP_Query arguments.
	 * @param string $attribute_to_count Attribute taxonomy name.
	 * @return array termId=>count pairs.
	 */
	public function get_attribute_counts( array $query_vars, string $attribute_to_count ) {
		/**
		 * Filter the data. @see get_filtered_price() for full documentation.
		 */
		$pre_filter_counts = apply_filters( 'woocommerce_pre_product_filter_data', null, 'attribute', $query_vars, array( 'taxonomy' => $attribute_to_count ) ); // phpcs:ignore WooCommerce.Commenting.CommentHooks.MissingSinceComment

if ( is_array( $pre_filter_counts ) ) {
			return $pre_filter_counts;
		}

$transient_key = $this->get_transient_key( $query_vars, 'attribute', array( 'taxonomy' => $attribute_to_count ) );
		$cached_data   = $this->get_cache( $transient_key );

if ( ! empty( $cached_data ) ) {
			return $cached_data;
		}

$results     = array();
		$product_ids = $this->get_cached_product_ids( $query_vars );

if ( $product_ids ) {
			global $wpdb;

// Optimization note: We evaluated using wc_product_attributes_lookup but decided against it, as removing
			// the posts table join in the query below produced better benchmarking results and required minimal changes.
			$taxonomy_escaped    = esc_sql( wc_sanitize_taxonomy_name( $attribute_to_count ) );
			$attribute_count_sql = "
				SELECT COUNT( DISTINCT term_relationships.object_id ) as term_count, terms.term_id as term_count_id
				FROM {$wpdb->term_relationships} AS term_relationships
				INNER JOIN {$wpdb->term_taxonomy} AS term_taxonomy USING( term_taxonomy_id )
				INNER JOIN {$wpdb->terms} AS terms USING( term_id )
				WHERE term_relationships.object_id IN ( {$product_ids} )
				AND term_taxonomy.taxonomy = '{$taxonomy_escaped}'
				GROUP BY terms.term_id
			";

/**
			 * We can't use $wpdb->prepare() here because using %s with
			 * $wpdb->prepare() for a subquery won't work as it will escape the
			 * SQL query.
			 * We're using the query as is, same as Core does.
			 */
			// phpcs:ignore WordPress.DB.PreparedSQL.NotPrepared
			$results = $wpdb->get_results( $attribute_count_sql );
			$results = array_map( 'absint', wp_list_pluck( $results, 'term_count', 'term_count_id' ) );
		}

/**
		 * Filter the results. @see get_filtered_price() for full documentation.
		 *
		 * @since 9.9.0
		 */
		$results = apply_filters( 'woocommerce_product_filter_data', $results, 'attribute', $query_vars, array( 'taxonomy' => $attribute_to_count ) );

$this->set_cache( $transient_key, $results );

return $results;
	}

/**
	 * Get taxonomy counts for the current products.
	 *
	 * @param array  $query_vars The WP_Query arguments.
	 * @param string $taxonomy_to_count   Taxonomy name.
	 * @return array termId=>count pairs.
	 */
	public function get_taxonomy_counts( array $query_vars, string $taxonomy_to_count ) {
		/**
		 * Filter the data. @see get_filtered_price() for full documentation.
		 *
		 * @since 9.9.0
		 */
		$pre_filter_counts = apply_filters( 'woocommerce_pre_product_filter_data', null, 'taxonomy', $query_vars, array( 'taxonomy' => $taxonomy_to_count ) );

if ( is_array( $pre_filter_counts ) ) {
			return $pre_filter_counts;
		}

$transient_key = $this->get_transient_key( $query_vars, 'taxonomy', array( 'taxonomy' => $taxonomy_to_count ) );
		$cached_data   = $this->get_cache( $transient_key );

if ( ! empty( $cached_data ) ) {
			return $cached_data;
		}

$results     = array();
		$product_ids = $this->get_cached_product_ids( $query_vars );

if ( $product_ids ) {
			global $wpdb;

$taxonomy_escaped = esc_sql( wc_sanitize_taxonomy_name( $taxonomy_to_count ) );

if ( is_taxonomy_hierarchical( $taxonomy_to_count ) ) {
				$results = $this->get_hierarchical_taxonomy_counts( $product_ids, $taxonomy_to_count );
			} else {
				$taxonomy_count_sql = "
					SELECT COUNT( DISTINCT term_relationships.object_id ) as term_count, term_taxonomy.term_taxonomy_id as term_count_id
					FROM {$wpdb->term_relationships} AS term_relationships
					INNER JOIN {$wpdb->term_taxonomy} AS term_taxonomy USING( term_taxonomy_id )
					WHERE term_relationships.object_id IN ( {$product_ids} )
					AND term_taxonomy.taxonomy = '{$taxonomy_escaped}'
					GROUP BY term_taxonomy.term_taxonomy_id
				";

/**
				 * We can't use $wpdb->prepare() here because using %s with
				 * $wpdb->prepare() for a subquery won't work as it will escape the
				 * SQL query.
				 * We're using the query as is, same as Core does.
				 */
				// phpcs:ignore WordPress.DB.PreparedSQL.NotPrepared
				$base_results = $wpdb->get_results( $taxonomy_count_sql );
				$results      = array_map( 'absint', wp_list_pluck( $base_results, 'term_count', 'term_count_id' ) );
			}
		}

/**
		 * Filter the results. @see get_filtered_price() for full documentation.
		 *
		 * @since 9.9.0
		 */
		$results = apply_filters( 'woocommerce_product_filter_data', $results, 'taxonomy', $query_vars, array( 'taxonomy' => $taxonomy_to_count ) );

$this->set_cache( $transient_key, $results );

return $results;
	}

/**
	 * Get hierarchical taxonomy counts using optimized hierarchy data.
	 *
	 * @param string $product_ids   Comma-separated list of product IDs.
	 * @param string $taxonomy_name Original taxonomy name for hierarchy methods.
	 * @return array Array of term_id => count pairs.
	 */
	private function get_hierarchical_taxonomy_counts( string $product_ids, string $taxonomy_name ) {
		global $wpdb;

// Step 1: Get all terms that have products in the filtered set (1 query).
		$taxonomy_escaped = esc_sql( wc_sanitize_taxonomy_name( $taxonomy_name ) );
		$base_terms_sql   = "
			SELECT DISTINCT tt.term_id, tt.term_taxonomy_id
			FROM {$wpdb->term_relationships} tr
			INNER JOIN {$wpdb->term_taxonomy} tt ON tr.term_taxonomy_id = tt.term_taxonomy_id
			WHERE tr.object_id IN ( {$product_ids} )
			AND tt.taxonomy = '{$taxonomy_escaped}'
		";

// phpcs:ignore WordPress.DB.PreparedSQL.NotPrepared
		$base_terms = $wpdb->get_results( $base_terms_sql );

if ( empty( $base_terms ) ) {
			return array();
		}

// Step 2: Build hierarchy relationships using TaxonomyHierarchyData.
		$hierarchy_counts = array();
		$processed_terms  = array();

// Process each base term and its ancestors.
		foreach ( $base_terms as $term ) {
			$term_id = (int) $term->term_id;

// Count for the term itself and all its descendants.
			if ( ! isset( $hierarchy_counts[ $term_id ] ) ) {
				$descendants                  = $this->taxonomy_hierarchy_data->get_descendants( $term_id, $taxonomy_name );
				$descendants[]                = $term_id; // Include the term itself.
				$hierarchy_counts[ $term_id ] = $descendants;
			}

// Get ancestors using hierarchy data.
			$ancestors = $this->taxonomy_hierarchy_data->get_ancestors( $term_id, $taxonomy_name );
			foreach ( $ancestors as $ancestor_id ) {
				if ( in_array( $ancestor_id, $processed_terms, true ) ) {
					continue;
				}

$descendants   = $this->taxonomy_hierarchy_data->get_descendants( $ancestor_id, $taxonomy_name );
				$descendants[] = $ancestor_id; // Include the ancestor term itself.

$hierarchy_counts[ $ancestor_id ] = $descendants;
				$processed_terms[]                = $ancestor_id;
			}
		}

if ( empty( $hierarchy_counts ) ) {
			return array();
		}

// Step 3: Execute batch counting using a single query with CASE statements.
		$count_cases = array();
		foreach ( $hierarchy_counts as $term_id => $term_ids ) {
			$term_ids_str  = implode( ',', array_map( 'absint', $term_ids ) );
			$count_cases[] = "COUNT(DISTINCT CASE WHEN tt.term_id IN ({$term_ids_str}) THEN tr.object_id END) as count_{$term_id}";
		}

$batch_count_sql = '
			SELECT ' . implode( ', ', $count_cases ) . "
			FROM {$wpdb->term_relationships} tr
			INNER JOIN {$wpdb->term_taxonomy} tt ON tr.term_taxonomy_id = tt.term_taxonomy_id
			WHERE tr.object_id IN ( {$product_ids} )
			AND tt.taxonomy = '{$taxonomy_escaped}'
		";

// phpcs:ignore WordPress.DB.PreparedSQL.NotPrepared
		$count_result = $wpdb->get_row( $batch_count_sql, ARRAY_A );

if ( empty( $count_result ) ) {
			return array();
		}

// Parse results back to term_id => count format.
		$final_counts = array();
		foreach ( $hierarchy_counts as $term_id => $term_ids ) {
			$count_key = "count_{$term_id}";
			if ( isset( $count_result[ $count_key ] ) && $count_result[ $count_key ] > 0 ) {
				$final_counts[ $term_id ] = absint( $count_result[ $count_key ] );
			}
		}

return $final_counts;
	}

/**
	 * Get filter data transient key.
	 *
	 * @param array  $query_vars   The query arguments to calculate the filter data.
	 * @param string $filter_type The type of filter. Accepts price|stock|rating|attribute.
	 * @param array  $extra        Some filter types require extra arguments for calculation, like attribute.
	 */
	private function get_transient_key( $query_vars, $filter_type, $extra = array() ) {
		return sprintf(
			'wc_%s_%s',
			CacheController::CACHE_GROUP,
			md5(
				wp_json_encode(
					array(
						'query_vars'  => $this->normalize_query_vars( $query_vars ),
						'extra'       => $extra,
						'filter_type' => $filter_type,
					)
				)
			)
		);
	}

/**
	 * Normalise query vars for cache key generation so that logically equivalent
	 * filter combinations produce the same hash.
	 *
	 * Rules applied (cache key only – the original $query_vars are never modified):
	 * - All keys are sorted alphabetically (ksort).
	 * - Values for keys that start with "filter_", equal "rating_filter", or are
	 *   built-in taxonomy short-names ("categories", "tags", "brands"):
	 *   comma-separated items are trimmed, lower-cased, sorted, then re-joined.
	 * - Values for keys that start with "query_type_": trimmed and lower-cased.
	 * - Values for "min_price" / "max_price": trimmed.
	 *
	 * @since 10.8.0
	 *
	 * @param array $query_vars Raw query vars.
	 * @return array Normalised copy of $query_vars.
	 */
	private function normalize_query_vars( array $query_vars ): array {
		// Built-in taxonomy filter params that are treated as unordered sets.
		// See Params::get_taxonomy_params() for the source of these short names.
		$taxonomy_set_params = array( 'categories', 'tags', 'brands' );

ksort( $query_vars );

foreach ( $query_vars as $key => $value ) {
			if ( ! is_string( $key ) || ! is_string( $value ) ) {
				continue;
			}

if ( str_starts_with( $key, 'filter_' ) || 'rating_filter' === $key || in_array( $key, $taxonomy_set_params, true ) ) {
				$pieces = array_map( 'trim', explode( ',', $value ) );
				$pieces = array_map( 'strtolower', $pieces );
				$pieces = array_values( array_unique( array_filter( $pieces, static fn( string $p ): bool => '' !== $p ) ) );
				sort( $pieces );
				$query_vars[ $key ] = implode( ',', $pieces );
			} elseif ( str_starts_with( $key, 'query_type_' ) ) {
				$query_vars[ $key ] = strtolower( trim( $value ) );
			} elseif ( 'min_price' === $key || 'max_price' === $key ) {
				$query_vars[ $key ] = trim( $value );
			}
		}

return $query_vars;
	}

/**
	 * Get cached filter data.
	 *
	 * @param string $key Transient key.
	 */
	private function get_cache( $key ) {
		if ( defined( 'WP_DEBUG' ) && WP_DEBUG ) {
			return null;
		}

$cache             = get_transient( $key );
		$transient_version = WC_Cache_Helper::get_transient_version( CacheController::CACHE_GROUP );

if ( empty( $cache['version'] ) ||
			! is_array( $cache['value'] ) ||
			empty( $cache['value'] ) ||
			$transient_version !== $cache['version']
		) {
			return null;
		}

return $cache['value'];
	}

/**
	 * Set the cache with transient version to invalidate all at once when needed.
	 *
	 * When the number of cached filter combinations reaches the configured
	 * maximum (default 1000), new combinations are silently skipped rather than
	 * stored, preventing unbounded transient growth from bot enumeration.
	 * The counter resets whenever the filter-data cache is invalidated.
	 * The limit can be adjusted via the `woocommerce_product_filter_cache_max_entries`
	 * filter. Set it to 0 to disable the cap entirely.
	 *
	 * @since 10.8.0 Cache-entry cap added.
	 *
	 * @param string $key   Transient key.
	 * @param mixed  $value Value to set.
	 *
	 * @return bool True if the cache was set, false otherwise.
	 */
	private function set_cache( $key, $value ) {
		if ( ! is_array( $value ) ) {
			return false;
		}

/**
		 * Maximum number of cache entries (not unique filter combos).
		 *
		 * Each unique query-vars combo can produce up to 5 entries (price,
		 * stock, rating, attribute, taxonomy), so the effective cap on
		 * unique combos is roughly max_entries / 5.
		 *
		 * When the limit is reached, new entries are skipped until the
		 * cache is next invalidated.  Set to 0 to disable the cap.
		 *
		 * @hook woocommerce_product_filter_cache_max_entries
		 * @since 10.8.0
		 *
		 * @param int $max_entries Maximum number of cache entries. Default 1000.
		 * @return int
		 */
		$max_entries = (int) apply_filters( 'woocommerce_product_filter_cache_max_entries', 1000 );

if ( $max_entries > 0 ) {
			$count = (int) get_transient( CacheController::CACHE_ENTRY_COUNT_TRANSIENT );

if ( $count >= $max_entries ) {
				return false;
			}

// The counter only increments — it does not decrement when entries
			// expire naturally. The effective cap may therefore be reached
			// before $max_entries live transients exist, making the limit
			// slightly conservative. This is intentional: accuracy here is
			// not worth the cost of tracking individual expirations.
			set_transient( CacheController::CACHE_ENTRY_COUNT_TRANSIENT, $count + 1, DAY_IN_SECONDS );
		}

$transient_version = WC_Cache_Helper::get_transient_version( CacheController::CACHE_GROUP );
		$transient_value   = array(
			'version' => $transient_version,
			'value'   => $value,
		);

return set_transient( $key, $transient_value, DAY_IN_SECONDS );
	}

/**
	 * Get cached product IDs from query vars.
	 *
	 * Executes a WP_Query with the given query vars and returns a comma-separated string of product IDs.
	 * Results are cached to avoid repeated database queries.
	 *
	 * @param array $query_vars The WP_Query arguments.
	 * @return string Comma-separated list of product IDs.
	 */
	private function get_cached_product_ids( array $query_vars ) {
		$cache_key = WC_Cache_Helper::get_cache_prefix( CacheController::CACHE_GROUP ) . md5( wp_json_encode( $this->normalize_query_vars( $query_vars ) ) );
		$cache     = wp_cache_get( $cache_key );

if ( $cache ) {
			return $cache;
		}

add_filter( 'posts_clauses', array( $this->query_clauses, 'add_query_clauses' ), 10, 2 );
		add_filter( 'posts_pre_query', '__return_empty_array' );

$query_vars['no_found_rows']  = true;
		$query_vars['posts_per_page'] = -1;
		$query_vars['fields']         = 'ids';
		$query                        = new \WP_Query();

$query->query( $query_vars );

remove_filter( 'posts_clauses', array( $this->query_clauses, 'add_query_clauses' ), 10 );
		remove_filter( 'posts_pre_query', '__return_empty_array' );

global $wpdb;

// The query is already prepared by WP_Query.
		$results = $wpdb->get_results( $query->request, ARRAY_A ); // phpcs:ignore WordPress.DB.PreparedSQL.NotPrepared

if ( ! $results ) {
			$results = array();
		}

$results = implode( ',', array_column( $results, 'ID' ) );

wp_cache_set( $cache_key, $results );

return $results;
	}
}