SQL Server的慢，從大表

Question

選擇我有2代真正的大SQL Server數據庫表的物聯網項目

第一個表是消息（行計數7423889085行）

CREATE TABLE [aymax].[Message](
    [MessageId] [bigint] IDENTITY(1,1) NOT NULL, 
    [ObjectId] [int] NOT NULL, 
    [TimeStamp] [datetime] NOT NULL CONSTRAINT [DF__Message__TimeSta__3B75D760] DEFAULT (getdate()), 
    [GpsTime] [datetime] NOT NULL, 
    [VisibleSatelites] [int] NOT NULL, 
    [X] [float] NOT NULL, 
    [Y] [float] NOT NULL, 
CONSTRAINT [Message_PK] PRIMARY KEY NONCLUSTERED 
(
    [MessageId] 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 

第二個表是MessageSensors，行數（26359568037行），該表具有值在消息表的每個傳感器

CREATE TABLE [aymax].[MessageSensors](
    [MessageId] [bigint] NOT NULL, 
    [DataSourceId] [int] NOT NULL, 
    [Value] [float] NOT NULL CONSTRAINT [DF__AnalogDat__Value__5812160E] DEFAULT ((0)), 
CONSTRAINT [AnalogData_PK] PRIMARY KEY CLUSTERED 
(
    [MessageId] ASC, 
    [DataSourceId] 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 

我的問題，即由時間間隔2日期時間之間求是很慢，一個如果我選擇與消息傳感器數據LSO它變得更慢，也當我使用sp_BlitzIndex檢查從brentozar.com它說我有

「Indexaphobia：高價值的缺失索引」

1. [aymax ] [MessageSensors]（等式：[DataSourceID的]，[值]，包括：[的MessageId]）
2. [aymax] [MessageSensors]平等：[數值]包括：[的MessageId]，[的DataSourceID]

我相信創造這個2指數是將增加存儲了很多，也將花費太多時間被創建，我需要你的意見針對表有關指數

我目前的指標

1-

CREATE NONCLUSTERED INDEX [IX_gpstime_objectid] ON [aymax].[Message] 
(
    [GpsTime] ASC 
) 
INCLUDE ( [MessageId], 
    [ObjectId]) WITH (PAD_INDEX = OFF, STATISTICS_NORECOMPUTE = OFF, SORT_IN_TEMPDB = OFF, DROP_EXISTING = OFF, ONLINE = OFF, ALLOW_ROW_LOCKS = ON, ALLOW_PAGE_LOCKS = ON) 
GO 

2-

alter TABLE [aymax].[Message] ADD CONSTRAINT [Message_PK] PRIMARY KEY NONCLUSTERED 
    (
     [MessageId] ASC 
    )WITH (PAD_INDEX = OFF, STATISTICS_NORECOMPUTE = OFF, SORT_IN_TEMPDB = OFF, IGNORE_DUP_KEY = OFF, ONLINE = OFF, ALLOW_ROW_LOCKS = ON, ALLOW_PAGE_LOCKS = ON) 
    GO 

3rd-

ALTER TABLE [aymax].[MessageSensors] ADD CONSTRAINT [AnalogData_PK] PRIMARY KEY CLUSTERED 
(
    [MessageId] ASC, 
    [DataSourceId] ASC 
)WITH (PAD_INDEX = OFF, STATISTICS_NORECOMPUTE = OFF, SORT_IN_TEMPDB = OFF, IGNORE_DUP_KEY = OFF, ONLINE = OFF, ALLOW_ROW_LOCKS = ON, ALLOW_PAGE_LOCKS = ON) 
GO 

4-

CREATE NONCLUSTERED INDEX [MessageData_DataSourceId_IDX] ON [aymax].[MessageSensors] 
(
    [DataSourceId] ASC 
)WITH (PAD_INDEX = OFF, STATISTICS_NORECOMPUTE = OFF, SORT_IN_TEMPDB = OFF, DROP_EXISTING = OFF, ONLINE = OFF, ALLOW_ROW_LOCKS = ON, ALLOW_PAGE_LOCKS = ON) 
GO 

任何幫助，請，我需要做一個快速的信息檢索和信息傳感器

更新 而做了一些調查，我發現選擇浮點值將放緩結果太多，從1秒到3分鐘

SELECT m.messageid, 
       m.objectid, 
       m.gpstime, 
       m.x, 
       m.y, 
       -- slow is here if i replace md.value with md.messageId will return fast , md.value is float 
       md.Value , 
       0 
     FROM aymax.[message] m WITH (nolock) 
       left JOIN aymax.MessageSensors md WITH (nolock) 
         ON m.messageid = md.messageid 
         AND md.datasourceid = 425732 


     WHERE m.objectid = 14099 
       AND m.gpstime BETWEEN '2017-04-01 19:46:18.607' AND '2017-04-10 19:05:18.607' 

Answer 1

可能的解決方案：

1. 已過濾索引（按日期過濾且不索引舊數據） https://docs.microsoft.com/en-us/sql/relational-databases/indexes/create-filtered-indexes。
2. GpsTime，MessageId上的聚集索引（特別是如果您沒有關於其他索引的計劃）。需要重建你的桌子。
3. 分區（請參閱@ Siyaul的評論）