如果資料列的匯總程度不足以保護使用者隱私,或不符合廣告資料中心的其他隱私權檢查標準,查詢結果會捨棄這些資料列。您可以在分析查詢中設定篩除資料列摘要,這樣就能傳回一行資料列,內含已捨棄資料列的常數及總和,以免資料總數 (例如曝光或點擊總數) 出現落差。對於透過使用者介面建立的新分析查詢,預設會啟用篩除資料列摘要。篩除資料列摘要不適用於目標對象查詢。如果是透過 API 建立的查詢,則預設不會啟用篩除資料列摘要。
[null,null,["上次更新時間:2025-03-06 (世界標準時間)。"],[[["Ads Data Hub uses a filtered row summary to prevent discrepancies in data totals by aggregating data from rows that would otherwise be dropped due to privacy checks."],["Filtered row summaries are enabled by default for new analysis queries created via the UI but must be manually enabled for queries created via the API."],["Aggregation requirements still apply to the filtered row summaries; if the summarized data doesn't meet the thresholds, it will also be filtered, and a warning message will appear in the job status."],["You can manually configure filtered row summaries to specify which columns are aggregated and how, or enable automatic configuration for UI-created queries."],["When exporting multiple tables, you can add a filtered row summary directly in the column options using SQL syntax to specify aggregation behavior."]]],["Filtered row summaries aggregate data from rows dropped due to privacy checks or insufficient aggregation, preventing data discrepancies. They include sums from filtered rows and constants for identification. Automatic mode configures these summaries by default in the UI, adjusting as the query changes. Manual mode allows custom configuration of columns to aggregate and constants to label the summary. When using multiple tables in a query, filtered row summaries can be defined directly within the table creation. Using noise injection is recommended instead of difference checks.\n"]]