Note
Databricks CLI 的使用受限於 Databricks 授權 和 Databricks 隱私權聲明,包括任何使用數據條款。
ai-search
Databricks CLI 中的指令群組包含管理 AI 搜尋端點與索引的指令。 AI 搜尋端點代表用於承載 AI 搜尋索引的計算資源。 請參閱 AI 搜尋。
Databricks AI-Search 建立-端點
建立 AI 搜尋端點。
databricks ai-search create-endpoint PARENT ENDPOINT_TYPE [flags]
Arguments
PARENT
這個端點將被建立的工作區。 格式:workspaces/{workspace_id}
ENDPOINT_TYPE
端點的類型。 支援的值:STANDARD、STORAGE_OPTIMIZED
選項
--budget-policy-id string
使用者選擇的端點預算政策 ID。
--endpoint-id string
使用者提供的端點簡短名稱。
--json JSON
內嵌的 JSON 字串或 @path 整合至包含要求內容的 JSON 檔案。
--name string
AI 搜尋端點名稱。
--replica-count int
終端所需的複本數量。
--target-qps int
目標 QPS(每秒查詢數)為終端點設。
--usage-policy-id string
使用政策 ID 套用到端點。
Databricks AI-搜尋建立-索引
建立 AI 搜尋索引。
databricks ai-search create-index PARENT PRIMARY_KEY INDEX_TYPE [flags]
Arguments
PARENT
這個索引將建立的端點。 格式:workspaces/{workspace_id}/endpoints/{endpoint_id}
PRIMARY_KEY
索引的主鍵。 設定在創造狀態,之後是不可改變的。
INDEX_TYPE
索引類型。 支援的值:DELTA_SYNC、DIRECT_ACCESS
選項
--index-id string
使用者提供的 Unity 目錄資料表名稱用於索引。
--index-subtype IndexSubtype
指數的子類型。 支援的值:FULL_TEXT、HYBRID、VECTOR
--json JSON
內嵌的 JSON 字串或 @path 整合至包含要求內容的 JSON 檔案。
--name string
AI 搜尋索引名稱。
Databricks AI-搜尋刪除-端點
刪除 AI 搜尋端點。
databricks ai-search delete-endpoint NAME [flags]
Arguments
NAME
要刪除端點的完整資源名稱。 格式:workspaces/{workspace_id}/endpoints/{endpoint_id}
選項
Databricks AI-搜尋刪除-索引
刪除 AI 搜尋索引。
databricks ai-search delete-index NAME [flags]
Arguments
NAME
要刪除索引的完整資源名稱。 格式:workspaces/{workspace_id}/endpoints/{endpoint_id}/indexes/{index_id}
選項
Databricks AI-Search Get-Endpoint
取得 AI 搜尋端點。
databricks ai-search get-endpoint NAME [flags]
Arguments
NAME
端點的完整資源名稱。 格式:workspaces/{workspace_id}/endpoints/{endpoint_id}
選項
Databricks AI-Search 取得-索引
取得 AI 搜尋索引。
databricks ai-search get-index NAME [flags]
Arguments
NAME
索引的完整資源名稱。 格式:workspaces/{workspace_id}/endpoints/{endpoint_id}/indexes/{index_id}
選項
Databricks AI-Search List -endpoints
列出 AI 搜尋端點。
databricks ai-search list-endpoints PARENT [flags]
Arguments
PARENT
擁有這組端點的工作空間。 格式:workspaces/{workspace_id}
選項
--limit int
要傳回的結果數目上限。
--page-size int
這是對結果回傳數量的最佳努力上限。
DataBricks AI-搜尋列表-索引
列出 AI 搜尋索引。
databricks ai-search list-indexes PARENT [flags]
Arguments
PARENT
擁有這組索引的端點。 格式:workspaces/{workspace_id}/endpoints/{endpoint_id}
選項
--limit int
要傳回的結果數目上限。
--page-size int
這是對結果回傳數量的最佳努力上限。
Databricks AI-搜尋查詢索引
查詢 AI 搜尋索引。
databricks ai-search query-index NAME [flags]
Arguments
NAME
查詢索引的完整資源名稱。 格式:workspaces/{workspace_id}/endpoints/{endpoint_id}/indexes/{index_id}
選項
--filters-json string
用 JSON 字串描述查詢過濾器。
--json JSON
內嵌的 JSON 字串或 @path 整合至包含要求內容的 JSON 檔案。
--max-results int
要傳回的結果數目上限。
--query-text string
查詢文字。
--query-type string
查詢類型: ANN、、 HYBRID或 FULL_TEXT。
--score-threshold float
近似最近鄰搜尋的分數閾值。
Databricks AI-搜尋移除-資料
從 AI 搜尋索引中移除資料。
databricks ai-search remove-data NAME [flags]
Arguments
NAME
索引的完整資源名稱。 一定是直接存取索引。 格式:workspaces/{workspace_id}/endpoints/{endpoint_id}/indexes/{index_id}
選項
--json JSON
內嵌的 JSON 字串或 @path 整合至包含要求內容的 JSON 檔案。
DataBricks AI-搜尋掃描-索引
掃描 AI 搜尋索引。
databricks ai-search scan-index NAME [flags]
Arguments
NAME
掃描所需的索引完整資源名稱。 格式:workspaces/{workspace_id}/endpoints/{endpoint_id}/indexes/{index_id}
選項
--json JSON
內嵌的 JSON 字串或 @path 整合至包含要求內容的 JSON 檔案。
--page-size int
本頁最多可返回的列數。
--page-token string
來自先前回覆的頁面標記。 若未設定,掃描則從頭開始。
Databricks AI-搜尋同步-索引
同步 AI 搜尋索引。
databricks ai-search sync-index NAME [flags]
Arguments
NAME
索引的完整資源名稱以同步。 一定是 Delta Sync 索引。 格式:workspaces/{workspace_id}/endpoints/{endpoint_id}/indexes/{index_id}
選項
Databricks AI-Search 更新-端點
更新 AI 搜尋端點。
databricks ai-search update-endpoint NAME UPDATE_MASK ENDPOINT_TYPE [flags]
Arguments
NAME
AI 搜尋端點名稱。 輸出的完整資源路徑(格式: workspaces/{workspace_id}/endpoints/{endpoint_id})。
UPDATE_MASK
要更新的欄位清單。
ENDPOINT_TYPE
端點的類型。 支援的值:STANDARD、STORAGE_OPTIMIZED
選項
--budget-policy-id string
使用者選擇的端點預算政策 ID。
--json JSON
內嵌的 JSON 字串或 @path 整合至包含要求內容的 JSON 檔案。
--name string
AI 搜尋端點名稱。
--replica-count int
終端所需的複本數量。
--target-qps int
目標 QPS(每秒查詢數)為終端點設。
--usage-policy-id string
使用政策 ID 套用到端點。
Databricks AI-Search upsert-data
將資料上傳到 AI 搜尋索引。
databricks ai-search upsert-data NAME INPUTS_JSON [flags]
Arguments
NAME
索引的完整資源名稱。 一定是直接存取索引。 格式:workspaces/{workspace_id}/endpoints/{endpoint_id}/indexes/{index_id}
INPUTS_JSON
JSON 文件描述要上傳的列數。
選項
--json JSON
內嵌的 JSON 字串或 @path 整合至包含要求內容的 JSON 檔案。
全域旗標
--debug
是否要啟用偵錯記錄。
-h 或 --help
顯示 Databricks CLI、相關命令群組或相關命令的幫助說明。
--log-file 字串
字串,表示要寫入輸出記錄檔的檔案。 若未指定此旗標,則預設值是將輸出記錄寫入 stderr。
--log-format 格式
記錄格式類型,text 或 json。 預設值為 text。
--log-level 字串
表示日誌格式層級的字串。 若未指定,則日誌的格式層級功能將被停用。
-o, --output 類型
指令輸出類型,text 或 json。 預設值為 text。
-p, --profile 字串
用於執行命令的~/.databrickscfg檔案中的配置檔名稱。 如果未指定此旗標,則如果存在,則會使用名為 DEFAULT 的配置檔。
--progress-format 格式
顯示進度記錄的格式:default、、appendinplace、 或json
-t, --target 字串
如果適用,要使用的套件組合目標