如何使用MS SQL將相同表中的父項記錄複製到子項CTE

Question

我有要求將同一父記錄複製到同一表的N個子項。 數據中的唯一區別是父母標識和客戶標識，我必須將父級記錄複製到只有ParentId和ClientId值更改的子級。 - 父記錄可能是1到2000，需要爲每個客戶端移動。 - 兒童記錄（重複1至1000）不同的孩子。

我已經實現了使用SQL Server Cursor的要求，但是如果我們有超過1000的父記錄和超過1500的子（重複條目），那麼這非常慢。所以我想知道是否可以使用CTE而是使用光標來獲得批量插入的性能。

我使用下面的代碼父記錄複製到每個客戶端：

DECLARE @id bigint; 
--Get list of client which will repeat the rows from Master entries 
--This will return three rows Id:1,5,7 
DECLARE client_cursor CURSOR FOR SELECT Id FROM Client WHERE ClientName like 'info%' 
OPEN client_cursor 
    FETCH NEXT FROM client_cursor INTO @id; 
    WHILE @@FETCH_STATUS = 0 
    BEGIN 
     /*....*/ 
     /* Other business logic performed here */ 
     /*....*/ 
     INSERT INTO Resources (ParentId,ClientId,Location,Title,[Status],Alert) 
     --This will return two rows Id:1,2 
     SELECT Id,@id,Location,Title,[Status],Alert FROM Resources WHERE ParentId IS NULL AND ClientId IS NULL 

     FETCH NEXT FROM client_cursor INTO @id 
    END 
CLOSE client_cursor; 
DEALLOCATE client_cursor; 

以下是之前和使用上面的代碼片斷添加增量家長記錄孩子後的屏幕截圖。

-- Table Creation 
CREATE TABLE [dbo].[Client](
    [Id] [bigint] IDENTITY(1,1) NOT NULL, 
    [ClientName] [varchar](50) NULL, 
CONSTRAINT [PK_Client] PRIMARY KEY CLUSTERED 
(
    [Id] ASC 
)WITH (PAD_INDEX = OFF, STATISTICS_NORECOMPUTE = OFF, IGNORE_DUP_KEY = OFF, ALLOW_ROW_LOCKS = ON, ALLOW_PAGE_LOCKS = ON) ON [PRIMARY] 
) ON [PRIMARY] 
GO 

CREATE TABLE [dbo].[Resources](
    [Id] [bigint] IDENTITY(1,1) NOT NULL, 
    [ParentId] [bigint] NULL, 
    [ClientId] [bigint] NULL, 
    [Location] [varchar](50) NULL, 
    [Title] [varchar](50) NULL, 
    [Status] [varchar](50) NULL, 
    [Alert] [varchar](50) NULL, 
CONSTRAINT [PK_Resources] PRIMARY KEY CLUSTERED 
(
    [Id] ASC 
)WITH (PAD_INDEX = OFF, STATISTICS_NORECOMPUTE = OFF, IGNORE_DUP_KEY = OFF, ALLOW_ROW_LOCKS = ON, ALLOW_PAGE_LOCKS = ON) ON [PRIMARY] 
) ON [PRIMARY] 
GO 

-- Default Data Insertion 
SET IDENTITY_INSERT [dbo].[Client] ON 
INSERT [dbo].[Client] ([Id], [ClientName]) VALUES (1, N'Infosstretch') 
INSERT [dbo].[Client] ([Id], [ClientName]) VALUES (2, N'Microsoft') 
INSERT [dbo].[Client] ([Id], [ClientName]) VALUES (3, N'Sun System') 
INSERT [dbo].[Client] ([Id], [ClientName]) VALUES (4, N'IBM') 
INSERT [dbo].[Client] ([Id], [ClientName]) VALUES (5, N'Infosys') 
INSERT [dbo].[Client] ([Id], [ClientName]) VALUES (6, N'TCS') 
INSERT [dbo].[Client] ([Id], [ClientName]) VALUES (7, N'Infomatica') 
SET IDENTITY_INSERT [dbo].[Client] OFF 

SET IDENTITY_INSERT [dbo].[Resources] ON 
INSERT [dbo].[Resources] ([Id], [ParentId], [ClientId], [Location], [Title], [Status], [Alert]) VALUES (1, NULL, NULL, N'India', N'Master A', N'New', N'Issue with location') 
INSERT [dbo].[Resources] ([Id], [ParentId], [ClientId], [Location], [Title], [Status], [Alert]) VALUES (2, NULL, NULL, N'Australia', N'Master B', N'Updated', N'No major issue') 
SET IDENTITY_INSERT [dbo].[Resources] OFF 

Answer 1

有可能是由您的業務規則的一些原因，是爲簡潔，刪除在這裏，但爲什麼不這樣呢？

INSERT #Resources 
SELECT 
    R.Id AS ParentId, 
    C.Id AS ClientId, 
    R.Location, 
    R.Title, 
    R.Status, 
    R.Alert 
FROM #Resources R 
    CROSS JOIN #Client C 
WHERE R.ParentId IS NULL AND R.ClientId IS NULL -- Root level 
    AND NOT EXISTS (SELECT * FROM #Resources WHERE ParentId = R.Id) -- Not already populated 
    -- AND /* Other business logic performed here */