使用 GoogleAdsService 回報串流
透過集合功能整理內容
你可以依據偏好儲存及分類內容。
如要擷取 Google Ads API 實體和報表資料，請使用下列任一方法：
這兩種方法的主要差異如下：
SearchStream對Search
Search 可以傳送多個分頁要求來下載完整報表，但 SearchStream 只會傳送單一要求，並與 Google Ads API 建立持續連線，不受報表大小影響。
如果是 SearchStream，資料封包會立即開始下載，整個結果會快取在資料緩衝區中。您的程式碼可以開始讀取緩衝資料，不必等待整個串流完成。
視應用程式而定，SearchStream 可省去要求 Search 回應中每個網頁所需的網路往返時間，因此與分頁功能相比，SearchStream 可提升效能，尤其是在處理較大的報表時。
範例
這個範例會查看包含 100,000 個資料列的報表。下表列出這兩種方法在會計上的差異。
|
SearchStream |
搜尋 |
| 頁面大小 |
不適用 |
每頁 10,000 列 |
| API 要求數 |
1 項要求 |
10 項要求 |
| API 回應數量 |
1 個連續串流 |
10 則回覆 |
在大多數情況下，我們建議使用 SearchStream，而非 Search，原因如下：
頻率限制
兩種方法的每日上限都必須遵守開發人員權杖的標準限制和存取層級。無論結果是否分頁或串流，單一查詢或報表都會計為一項作業。
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：2025-08-27 (世界標準時間)。
[null,null,["上次更新時間：2025-08-27 (世界標準時間)。"],[[["\u003cp\u003eThe Google Ads API offers two primary methods for retrieving data: \u003ccode\u003eSearchStream\u003c/code\u003e and \u003ccode\u003eSearch\u003c/code\u003e, both suitable for production environments and fetching objects and reports.\u003c/p\u003e\n"],["\u003cp\u003e\u003ccode\u003eSearchStream\u003c/code\u003e delivers results as a continuous stream, ideal for large reports, while \u003ccode\u003eSearch\u003c/code\u003e provides paginated responses.\u003c/p\u003e\n"],["\u003cp\u003e\u003ccode\u003eSearchStream\u003c/code\u003e often offers performance advantages by reducing network round trips, especially for reports exceeding 10,000 rows.\u003c/p\u003e\n"],["\u003cp\u003eBoth methods are subject to the same daily limits and access levels, with a single query or report counting as one operation regardless of the method.\u003c/p\u003e\n"],["\u003cp\u003eGoogle recommends using \u003ccode\u003eSearchStream\u003c/code\u003e for most use cases due to its performance benefits for larger reports.\u003c/p\u003e\n"]]],[],null,["# Report streaming using GoogleAdsService\n\n\u003cbr /\u003e\n\n\u003cbr /\u003e\n\nTo retrieve Google Ads API entities and reporting data, use one of these methods:\n\n- [`GoogleAdsService.SearchStream`](/google-ads/api/reference/rpc/v21/GoogleAdsService/SearchStream)\n- [`GoogleAdsService.Search`](/google-ads/api/reference/rpc/v21/GoogleAdsService/Search)\n\nHere are the high-level distinctions for the two methods:\n\n| | GoogleAdsService.SearchStream | GoogleAdsService.Search |\n| Suitable for production code | **Yes** | **Yes** |\n| Service | [`GoogleAdsService`](/google-ads/api/reference/rpc/v21/GoogleAdsService) | [`GoogleAdsService`](/google-ads/api/reference/rpc/v21/GoogleAdsService) |\n| Scenario | Fetching objects and reports | Fetching objects and reports |\n| Response | **Stream** of [`GoogleAdsRow`](/google-ads/api/reference/rpc/v21/GoogleAdsRow) objects | Pages of [`GoogleAdsRow`](/google-ads/api/reference/rpc/v21/GoogleAdsRow) objects |\n| Response's fields | Only those specified in the query | Only those specified in the query |\n| Daily limits | Daily limits based on [access levels](/google-ads/api/docs/access-levels) | Daily limits based on [access levels](/google-ads/api/docs/access-levels) |\n|------------------------------|----------------------------------------------------------------------------------------|-----------------------------------------------------------------------------------|\n\n`SearchStream` versus `Search`\n------------------------------\n\nWhile [`Search`](/google-ads/api/reference/rpc/v21/GoogleAdsService/Search) can send multiple\npaginated requests to download the entire report, [`SearchStream`](/google-ads/api/reference/rpc/v21/GoogleAdsService/SearchStream) sends a single request and\ninitiates a persistent connection with the Google Ads API regardless of report size.\n\nFor `SearchStream`, data packets start to download immediately with the entire\nresult cached in a data buffer. Your code can start reading the buffered data\nwithout having to wait for the entire stream to finish.\n\nBy eliminating the round-trip network time required to request each individual\npage of a `Search` response, depending on your app, `SearchStream` can offer\nimproved performance over paging, especially for bigger reports.\n\n### Example\n\nThis example looks at a report that consists of `100,000` rows. The following\ntable breaks down the accounting differences between the two methods.\n\n| | SearchStream | Search |\n| Page size | Not Applicable | 10,000 rows per page |\n| Number of API requests | 1 request | 10 requests |\n| Number of API responses | 1 continuous stream | 10 responses |\n|-------------------------|---------------------|----------------------|\n\n### Performance factors\n\nFor most use cases, we recommend `SearchStream` over `Search` for the following\nreasons:\n\n- For single page reports (under 10,000 rows): No significant performance\n differences between the two methods.\n\n- For multiple page reports: `SearchStream` is typically faster since multiple\n round trips are avoided, and reading or writing from disk cache is less of a\n factor.\n\n### Rate limits\n\nDaily limits for both methods adhere to the standard limits and [access\nlevels](/google-ads/api/docs/access-levels) of your developer token. A single query or report\nis counted as one operation regardless of the result being paged or streamed."]]
