SQL Server performance deterioration -
I am currently working with the performance / memory consumption optimization of our applications. One of the executable tasks is to replace all the blobs in a table that corresponds to the zero values for the empty array; This will reduce the DB size, memory consumption and speed up the load. The definition of the table is:
Make Table [DBO]. [Sanpltebl] ([id] [Binint] not, tap [creation time] [day and time] NULL, [Bainreedeta] [image] with a tap (Piadi_aiandiaks = off, Stetsiks_anoarisiompiuti = off, Aijianoaraidiupi_kea = off, ALLOW_ROW_LOCKS = ON , ALLOW_PAGE_LOCKS =) [Avenue] [bit] nULL, [last savate Time] [Dettai] taps, Constrakt [PK-Simenttebl] primary cluster ([ID] ASC) [primary]) [primary] on
I updated the table and the image field values (bines Eedeta) updated with NULL values, where appropriate (data related to array empty app). Now, let me look at the performance poorly, when selecting * from the same selection. *
Originally those fields were updated which was length = 512 bytes, not sure if it counts.
Any idea why choose Blobs tap value takes more time to choose the actual binary data even if the data is the same for different lines?
I do not know the answer to this question. I though tried to do the following tests and as a result I was surprised.
Make Table [DBO]. [Sample table] ([id] [Begin] NOT NULL, [creation time] [DATETIME] zero, [binaryData] [IMAGE] zero [IsEvent] [BIT] zero [LastSavedTime] [DATETIME] zero barrier [PK_SampleTable ] In the primary key cluster ([id] ASC)) [dbo]. Select [SampleTable] 1, GETDATE (), 0x1111, 1, GETDATE () [dbo]. Select [SampleTable] 2, GETDATE (), 0x2222, 2, GETDATE () [dbo]. [SampleTable] Select 3, GETDATE (), Zero, 3, GETDATE () Update [dbo]. [SampleTable] set [binaryData] = zero where [id] = 2
I cast as the view was surprised to see this, I NULL
The difference between the row and the updated line on NULL
by me
It seems that whenever the value is updated to tap, just because of some reason zero Bitmap does not determine and still need to follow the indicator for another LOB_DATA
P age
inserted as zero
Updated to zero
Comments
Post a Comment