分享方式:


Marketplace 系統數據表參考

重要

這項功能處於公開預覽狀態

本文提供如何使用系統數據表來協助運作 Databricks Marketplace 銷售程式的概觀。

Marketplace 系統數據表存在於架構中 system.marketplace 。 下表可供使用:

  • 清單存取:記錄已完成 要求數據的 取用者資訊,或 取得清單上的數據 事件。
  • 漏鬥圖事件:記錄對清單所採取的印象和動作。

注意

您可以使用提供者分析儀錶板來監視收件者使用計量。 儀錶板會從 Marketplace 系統數據表提取數據。 請參閱 使用儀錶板監視列出使用計量。

列出存取事件數據表

數據表 listing_access_events 會擷取已完成 要求數據的 取用者資訊,或 取得清單上的數據 事件。

system.marketplace.listing_access_events 具有下列架構:

資料行名稱 資料類型 描述
account_id 字串 裝載清單的帳戶標識碼。
metastore_id 字串 裝載清單的符合項目識別碼。
metastore_cloud 字串 裝載清單之中繼存放區的雲端提供者。
metastore_region 字串 裝載清單的中繼存放區區域。
provider_id 字串 提供者配置檔識別碼。
provider_name 字串 提供者配置檔名稱。
listing_id 字串 清單標識碼。
listing_name 字串 清單名稱。
consumer_delta_sharing_recipient_name 字串 取用者的基礎 Delta 共用收件者名稱。 值為 null 當event_type為 REQUEST_DATA時。
consumer_delta_sharing_recipient_type 字串 取用者是否位於 Azure Databricks 帳戶上。 值會是 OPENDATABRICKS
consumer_cloud 字串 取用者的雲端。 如果 consumer_delta_sharing_recipient_type 為 ,則為 OPENNull。
consumer_region 字串 取用者的區域。 如果 consumer_delta_sharing_recipient_type 為 ,則為 OPENNull。
consumer_metastore_id 字串 取用者的中繼存放區標識碼。 如果 consumer_delta_sharing_recipient_type 為 ,則為 OPENNull。
consumer_email 字串 取用者的電子郵件位址。 PII。
consumer_name 字串 取用者的名稱。 PII。
consumer_company 字串 消費者的公司。
consumer_intended_use 字串 取用者的預定清單用途。
consumer_comments 字串 取用者留下的任何其他批注。
event_type 字串 存取的類型。 這個值可以是 REQUEST_DATAGET_DATA
event_date date 事件發生的 UTC 日期。
event_time timestamp 事件發生時的確切 UTC 時間戳。

列出漏鬥圖事件數據表

數據表 listing_funnel_events 會分析對清單所採取的印象和動作。 event_type會告訴您取用者對清單採取什麼動作。

system.marketplace.listing_funnel_events 具有下列架構:

資料行名稱 資料類型 描述
account_id 字串 裝載清單的帳戶標識碼。
metastore_id 字串 裝載清單的符合項目識別碼。
metastore_cloud 字串 裝載清單之中繼存放區的雲端提供者。
metastore_region 字串 裝載清單的中繼存放區區域。
provider_id 字串 提供者配置檔識別碼。
provider_name 字串 提供者配置檔名稱。
listing_id 字串 清單標識碼。
listing_name 字串 清單名稱。
event_type 字串 取用者動作的類型。 請參閱 事件類型
event_time timestamp 事件發生時的確切 UTC 時間戳。
event_date date 事件發生的 UTC 日期。
consumer_cloud 字串 取用者的雲端。 如果 consumer_delta_sharing_recipient_type 為 ,則為 OPENNull。
consumer_region 字串 取用者的區域。 如果 consumer_delta_sharing_recipient_type 為 ,則為 OPENNull。

事件類型

event_type資料列的可能值為:

  • VIEW_LISTING
  • START_REQUEST_DATA
  • COMPLETE_REQUEST_DATA
  • ABANDON_REQUEST_DATA
  • START_GET_DATA
  • COMPLETE_GET_DATA
  • ABANDON_GET_DATA
  • UNINSTALL_DATA
  • FAIL_REQUEST_DATA
  • FAIL_GET_DATA

已知的限制

  • 自 2024 年 3 月 12 日起,Marketplace 系統數據表包括使用 Google Cloud 上的 Databricks 取用者的活動。 數據表不包含此日期之前的 Google Cloud 取用者活動。
  • 在數據表中listing_funnel_events,2023 年 11 月 30 日之前未記錄 和 FAIL_GET_DATA 的事件類型ABANDON_FAIL_DATA

範例查詢

本節包含下列範例查詢,可用來深入了解清單上的取用者活動。

過去10天內的取用者要求

SELECT event_date, provider_name, listing_name, listing_id, consumer_delta_sharing_recipient_name, consumer_cloud, consumer_region, consumer_name, consumer_email, consumer_company
FROM system.marketplace.listing_access_events
WHERE event_type = 'REQUEST_DATA'
AND event_date >= date_add(current_date(), -10)

依要求數目列出的最上層清單

SELECT listing_name, consumer_cloud, count(*) as requestCount
FROM system.marketplace.listing_access_events
GROUP BY listing_name, consumer_cloud
ORDER BY requestCount DESC

檢視最上層要求者

SELECT consumer_name, consumer_email, count(*) as requestCount
FROM system.marketplace.listing_access_events
GROUP BY consumer_name, consumer_email
ORDER BY requestCount DESC

檢視清單上每個取用者動作的次數

SELECT event_type, COUNT(*) as occurrences
FROM system.marketplace.listing_funnel_events
WHERE listing_name = `{{listing_name}}`
GROUP BY event_type

筆記本示範:分析 Marketplace 系統數據表

下列筆記本包含本文中的範例查詢,以及提供進一步分析的其他數據格。

Marketplace 系統數據表示範

取得筆記本