存取實體讀取檔案
透過集合功能整理內容
你可以依據偏好儲存及分類內容。
實體讀取檔案是合作夥伴廣告活動物件的 JSON 表示法。實體讀取檔案為唯讀，無法用於將變更寫回 Google Display & Video 360。如要進一步瞭解如何使用實體讀取檔案，請與客戶經理聯絡，他們會進一步討論您的需求。
請注意，實體讀取檔案只會保留至處理當日起算的 60 天。
您可以在檔案格式參考資料中查看實體讀取檔案的格式。
實體讀取檔案會儲存為兩個群組：
- 公開資料表，內含常見的公開資料，例如
gdbm-public 值區中儲存的瀏覽器版本或語言相關資訊。
- 私人表格，其中包含合作夥伴的廣告活動、預算和其他私人資料的資訊。私人資料表會儲存在特定合作夥伴的值區中。舉例來說，如果合作夥伴 ID 為 123，則包含私人資料表的值區會是
gdbm-123。
公開和私人資料表都位於資料桶內的
entity 子目錄中。
如要進一步瞭解如何處理子目錄，請參閱
Cloud Storage 物件命名空間。
公開資料表會在世界標準時間上午 5 點產生。私人資料表會在合作夥伴時區的凌晨 5 點產生 (但在少數情況下可能會延遲)，並自動安全地儲存至
Google Cloud Storage。系統會在產生所有公開和私人資料表後，產生摘要檔案。
公開資料表
gdbm-public/entity/(processing_date).(batch_index).(entity_type).json
| 欄位 | 說明 |
|---|
processing_date | Display & Video 360 處理資料的日期，不一定與檔案內含的事件相關。日期格式為 YYYYMMDD，例如 20120320 代表 2012 年 3 月 20 日。 |
batch_index |
批次的索引，從零開始。舉例來說，如果每小時產生一次，第一批會是 0，第二批會是 1，以此類推。 |
entity_type | 檔案類型：SupportedExchange、DataPartner、UniversalSite、GeoLocation、Language、OperatingSystem、Browser、Isp、summary 和 SiteIdToPlacementId.csv。 |
舉例來說，如果您要產生檔案名稱，以便從 2013 年 1 月 15 日開始擷取
browser 資料表的實體讀取檔案，檔案路徑會像這樣：
gdbm-public/entity/20130115.0.Browser.json
私人表格
gdbm-(partner_id)/entity/(processing_date).(batch_index).(entity_type).json
| 欄位 | 說明 |
|---|
partner_id | 合作夥伴的數字 ID。 |
processing_date | Display & Video 360 處理資料的日期。這個日期不一定與檔案內的事件相關。日期格式為 YYYYMMDD，例如 20120320 代表 2012 年 3 月 20 日。系統將在合作夥伴時區的凌晨 4 點到上午 6 點之間產生檔案。 |
batch_index | 批次索引，從零開始。目前這項值一律為 0。 |
entity_type | 檔案類型：Partner、Advertiser、InsertionOrder、LineItem、Creative、Pixel、InventorySource、UserList、UniversalChannel 和 summary。 |
舉例來說，如果您要產生檔案名稱，以便擷取合作夥伴 123 自 2013 年 4 月 2 日起存取
line_item 資料表的實體讀取檔案，檔案名稱會類似以下格式：
gdbm-123/entity/20130402.0.LineItem.json
除非另有註明，否則本頁面中的內容是採用創用 CC 姓名標示 4.0 授權，程式碼範例則為阿帕契 2.0 授權。詳情請參閱《Google Developers 網站政策》。Java 是 Oracle 和/或其關聯企業的註冊商標。
上次更新時間：2025-08-31 (世界標準時間)。
[null,null,["上次更新時間：2025-08-31 (世界標準時間)。"],[[["\u003cp\u003eEntity Read Files (ERF) have been discontinued and users should migrate to the Display & Video 360 API.\u003c/p\u003e\n"],["\u003cp\u003eEntity Read Files provided read-only JSON representations of a partner's campaign data but were not capable of writing back changes.\u003c/p\u003e\n"],["\u003cp\u003eThese files were organized into public tables, containing common data like browser versions, and private tables, containing partner-specific campaign data, within respective Google Cloud Storage buckets.\u003c/p\u003e\n"],["\u003cp\u003ePublic tables were generated daily at 5 AM UTC, while private tables were generated at 5 AM in the partner's timezone and stored with a specific naming convention indicating the processing date, batch index, and entity type.\u003c/p\u003e\n"],["\u003cp\u003eEach public and private table has a \u003ccode\u003eprocessing_date\u003c/code\u003e and \u003ccode\u003ebatch_index\u003c/code\u003e as well as an \u003ccode\u003eentity_type\u003c/code\u003e to indicate what type of data is contained within the table.\u003c/p\u003e\n"]]],[],null,["# Access Entity Read Files\n\n| Entity Read Files have sunset. [Migrate to use the\n| Display \\& Video 360 API.](/display-video/api/guides/how-tos/erf-migration)\n\nEntity Read Files are JSON representations of a partner's campaign objects. Entity Read Files\nare read-only and cannot be used to write-back changes to Google Display \\& Video 360. To learn more about using\nEntity Read Files, contact your account manager, who will be able to discuss your needs in more detail.\n\nNote that Entity Read Files are available only for 60 days from their processing date.\n\nThe format of the entity read files can be viewed in the\n[File Format reference](/bid-manager/guides/entity-read/format-v2).\n\nThe entity read files are stored as two groups: \n\n- Public tables, which contain common public data, such as information about browser versions or languages that are stored in the `gdbm-public` bucket.\n- Private tables, which contain information about the partner's campaigns, budgets, and other private data. Private tables are stored in partner-specific buckets; for example, if your partner ID is 123, the bucket that contained your private tables would be `gdbm-123`.\n\nBoth public and private tables are in an `entity` subdirectory within the bucket. For more information about handling subdirectories, see [Cloud Storage object namespace](/storage/docs/objects#namespace). \nThe public tables are generated at 5AM UTC. The private tables are generated at 5AM in your partner's timezone (although in rare cases there may be delays) and stored automatically and securely in [Google Cloud Storage](/storage). A summary file is generated after all of the public and private tables have been generated. All filenames are case-sensitive.\n\nPublic Tables\n-------------\n\n```\ngdbm-public/entity/(processing_date).(batch_index).(entity_type).json\n```\n\n| Field | Description |\n|-------------------|----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|\n| `processing_date` | The date when Display \\& Video 360 processed the data and is not necessarily related to the events contained within the file. Date format is `YYYYMMDD`, for example `20120320` would be March 20, 2012. |\n| `batch_index` | A zero-based index of the batch. For example, if generated hourly the first batch would be `0`, the second batch would be `1`, and so on. |\n| `entity_type` | The type of file: `SupportedExchange`, `DataPartner`, `UniversalSite`, `GeoLocation`, `Language`, `OperatingSystem`, `Browser`, `Isp`, `summary` and `SiteIdToPlacementId.csv`. |\n\nFor example, if you were generating a filename to retrieve the entity read file for the `browser` table from January 15th, 2013, the path to the file would look like \n\n```\ngdbm-public/entity/20130115.0.Browser.json\n```\n\nPrivate Tables\n--------------\n\n```\ngdbm-(partner_id)/entity/(processing_date).(batch_index).(entity_type).json\n```\n\n| Field | Description |\n|-------------------|--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|\n| `partner_id` | The numeric ID of your Partner. |\n| `processing_date` | The date when Display \\& Video 360 processed the data. This date is not necessarily related to the events contained within the file. Date format is `YYYYMMDD`, for example `20120320` would be March 20, 2012. The files will be generated between 4AM and 6AM in the Partner's timezone. |\n| `batch_index` | A zero-based index of the batch. Currently this is always 0. |\n| `entity_type` | The type of file: `Partner`, `Advertiser`, `InsertionOrder`, `LineItem`, `Creative`, `Pixel`, `InventorySource`, `UserList`, `UniversalChannel`, and `summary`. |\n\nFor example, if you were generating a file name to retrieve the entity read file for partner 123 accessing the `line_item` table from April 2, 2013, your filename would look something like this: \n\n```\ngdbm-123/entity/20130402.0.LineItem.json\n```"]]
