Question

我在数据库中有两个表：tb_CampaignLead和tb_Feed。两个表都有数千行，主键上各有一个索引。 tb_CampaignLead的数据比tb_Feed多20％。对于tb_Feed，tb_CampaignLead仍然比SELECT *慢6倍。任何人都可以告诉我为什么tb_Feed这么慢。表结构如下：

CREATE TABLE [dbo].[tb_Feed](
    [FeedID] [uniqueidentifier] NOT NULL,
    [DateCreated] [smalldatetime] NULL,
    [LeadSourceID] [int] NULL,
    [RawLeadURL] [nvarchar](max) NULL,
    [CostPerEnquiry] [money] NULL,
    [ResultText] [nvarchar](max) NULL,
    [SrcResultText] [nvarchar](max) NULL,
    [SrcResult] [bit] NULL,
    [Encrypted] [bit] NULL,
    [ProductID] [int] NULL,
 CONSTRAINT [PK_tb_Feed] PRIMARY KEY CLUSTERED 


CREATE TABLE [dbo].[tb_CampaignLead](
    [LeadID] [uniqueidentifier] NOT NULL,
    [FeedID] [uniqueidentifier] NOT NULL,
    [CampaignID] [int] NOT NULL,
    [ProductID] [int] NOT NULL,
    [CompanyID] [int] NOT NULL,
    [LeadSourceID] [int] NOT NULL,
    [DateCreated] [smalldatetime] NULL,
    [LeadTextEmail] [nvarchar](max) NULL,
    [LeadTextOther] [nvarchar](max) NULL,
    [DateSent] [smalldatetime] NULL,
    [EmailResponse] [nvarchar](max) NULL,
    [OtherResponse] [nvarchar](max) NULL,
    [EmailOK] [bit] NULL,
    [OtherOK] [bit] NULL,
    [ResultPass] [bit] NULL,
    [L_Title] [nvarchar](50) NULL,
    [L_Email] [nvarchar](100) NULL,
    [L_Firstname] [nvarchar](max) NULL,
    [L_Surname] [nvarchar](max) NULL,
    [L_Address1] [nvarchar](max) NULL,
    [L_Address2] [nvarchar](max) NULL,
    [L_Address3] [nvarchar](max) NULL,
    [L_TownCity] [nvarchar](max) NULL,
    [L_AreaRegion] [nvarchar](max) NULL,
    [L_Country] [nvarchar](max) NULL,
    [L_PostCode] [nvarchar](50) NULL,
    [L_Telephone1] [nvarchar](50) NULL,
    [L_Telephone2] [nvarchar](50) NULL,
    [L_DOB] [smalldatetime] NULL,
    [L_Extra1] [nvarchar](max) NULL,
    [L_Extra2] [nvarchar](max) NULL,
    [L_Extra3] [nvarchar](max) NULL,
    [L_Extra4] [nvarchar](max) NULL,
    [L_Extra5] [nvarchar](max) NULL,
    [L_Extra6] [nvarchar](max) NULL,
    [L_Extra7] [nvarchar](max) NULL,
    [L_Extra8] [nvarchar](max) NULL,
    [L_Extra9] [nvarchar](max) NULL,
    [L_Extra10] [nvarchar](max) NULL,
    [L_Extra11] [nvarchar](max) NULL,
    [L_Extra12] [nvarchar](max) NULL,
    [L_Extra13] [nvarchar](max) NULL,
    [L_Extra14] [nvarchar](max) NULL,
    [L_Extra15] [nvarchar](max) NULL,
    [L_Extra16] [nvarchar](max) NULL,
    [L_Extra17] [nvarchar](max) NULL,
    [L_Extra18] [nvarchar](max) NULL,
    [L_Extra19] [nvarchar](max) NULL,
    [L_Extra20] [nvarchar](max) NULL,
    [SourceCost] [money] NULL,
    [CampaignCost] [money] NULL,
    [DeliveredPass] [bit] NULL,
    [FieldReqBWAND] [int] NULL,
    [FieldSuppliedBWAND] [int] NULL,
    [FilterBWAND] [int] NULL,
    [FilterPassBWAND] [int] NULL,
    [OPFilterBWAND] [int] NULL,
    [OPFilterPassBWAND] [int] NULL,
    [ProcessBWAND] [int] NULL,
    [ProcessPassBWAND] [int] NULL,
    [L_MobileNetwork] [nvarchar](max) NULL,
    [SrcResultPass] [bit] NULL,
 CONSTRAINT [PK_tb_CampaignLead] PRIMARY KEY CLUSTERED

Answer 1

使用评论中的这些信息：

campaignlead表的数据空间为734,832，供稿表的数据空间为2,595,792

如您所见，尽管行数较少，但tb_Feed的总数据量是tb_CampaignLead的3.5倍。当您在SSMS中执行SELECT * From ...时，它实际上必须检索那么多数据，格式化它，推送它通过网络连接（如果您从客户端运行），然后将其呈现到输出网格中。 / p>

对于小型数据集，这是如此之快，以至于它不是查询执行时间的重要部分。此外，它可以缓冲，使其看起来更快。但是，对于大型数据集，这可能需要相当长的时间，并且输出缓冲区/管道将开始备份。这不仅可以减慢查询执行时间，还可以成为其主导因素。

因此，对于SSMS中此大小的查询，我们实际上希望tb_Feed比tb_CampaignLead长约3.5倍。所以，虽然不完全一样，但你的结果肯定是在正确的球场。

这些表中还有其他一些因素可能导致额外的差异，包括每个表中“列外”存储了多少NVARCHAR(MAX)个数据。这是另一个对此有一些答案的问题：Should I use an inline varchar(max) column or store it in a separate table?

数据库表慢

1 个答案: