Skip to main content

What is Fill Factor in SQL Server

Fill Factor works in performance tuning area, for the index the most important property is Fill Factor. Fill Factor responsible to determine the percentage of space on each leaf-level page to be filled with data. As we know page is smallest unit of SQL server which size is 8k. Every page cans one or more than one row which is depending on size of row.

The Fill Factor specifies the % of fullness of the leaf level pages of an index. When an index is created or rebuild then filled up pages with data depend on Fill Factor. For example if we create an index and put the  value of Fill Factor is 70 then pages will filled up with data 70% other 30% space will be remain.

For Example, I am creating a Temp named table for testing of Index with Fill Factor.


CREATE TABLE Temp
(
       id INT IDENTITY(1,1),
       Name VARCHAR(100)
)

DECLARE @count INT=100000;
WHILE (@count>0)
BEGIN
       INSERT INTO Temp
       VALUES('SQL Server tutorial by codefari.com Type'+CONVERT(VARCHAR(100),@count))
       SET @count=@count-1
END

SELECT COUNT(*) FROM Temp




Now run following script


EXEC sp_spaceused 'dbo.Temp'



Result Set





High Fill Factor value

You can see index size 8kb and unused 8kb both are same, because we did not create any index on this table.

Now see following scrip I am creating a non-clustered Index with Fill Factor 100%




USE [Test]
GO

/****** Object:  Index [Name]    Script Date: 11/27/2015 6:34:53 PM ******/
CREATE NONCLUSTERED INDEX [Name] ON [dbo].[Temp]
(
       [Name] 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,
 FILLFACTOR = 100) ON [PRIMARY]
GO




Again run following script



EXEC sp_spaceused 'dbo.Temp'




Result Set






Here Index size become 6404 kb and unused 176kb. Here we have taken Fill Factor 100%. 
Note: You may choose high Fill Factor value if there is very little or no changes the underlying table's data. Means if you have an index that is constantly changing you would want to have a lower value to keep some free space available for new index entries.  Otherwise SQL Server would have to constantly do page splits to fit the new values into the index pages.

Low Fill Factor value
Now if we put value of Fill Factor as 50% then what will happen you may see in following example.
Drop the created index first and run the following script again.

USE [Test]
GO

/****** Object:  Index [Name]    Script Date: 11/27/2015 6:34:53 PM ******/
CREATE NONCLUSTERED INDEX [Name] ON [dbo].[Temp]
(
       [Name] 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,
 FILLFACTOR = 50) ON [PRIMARY]
GO


Again run the following script

EXEC sp_spaceused 'dbo.Temp'


Result Set





Now here you can see index_size is 12544kb and unused 248kb.



Note: With new data records added, the index pages need to have sufficient space to take the new entries. When there is not enough space a page split needs to occur which could impact performance depending on how frequently page splits need to occur.
 


Popular posts from this blog

Merge and Merge join transformation in SSIS

MERGE TRANSFORMATION
Using Merge Transformation we can combine two sorted data-set into single data-set basically Merge Transformation used to combines rows from two sorted data flows into one sorted data flow. Following tasks you may perform using Merge Transformation: 1.Suppose we have a scenario like, we need to merge data from a database table and excel means we want to merge data from two different data sources. For such type of scenario, you can use Merge Transformation. 2.If we want to merge data from two same structured tables but exists two different servers. 3.Sometimes we get an error due to data in a row, after correcting errors in the data we can re-merge rows easily. See below explanations may help you to understand Merge Transformation: I do evaluate here, you already know about the data source, data conversion, data flow, task flow, control flow etc. Note:Before Merge transformation, we need to sort the data using Sort Transformation. After sorting data add data path to Merge…

Add day to ISODate in MongoDB

We can use $add operator to add days in ISODate in mongodb, $add is the Arithmetic Aggregation Operator which adds number and date in mongodb.
Syntax:

{ $add: [ <expression1>, <expression2>, ... ] }

Note:  If one of the argument is date $add operator treats to other arguments as milliseconds to add to the date.
Example: Suppose we have a Test collection as below.

{"Title" : "Add day to ISODate in MongoBD","CreatedDate" : ISODate("2016-07-07T08:00:00.000Z")}

Query to add 2 days in CreatedDate

db.Test.aggregate([      { $project: { Title: 1, AddedDate: { $add: [ "$CreatedDate", 2*24*60*60000 ] } } }    ])

Result:

{ "_id" : ObjectId("579a1567ac1b3f3732483de0"), "Title" : "Add day to ISODate in MongoBD", "AddedDate" : ISODate("2016-07-09T08:00:00.000Z") }

Note: As mentioned in above note we have to convert days in millisecond because $add operator treat to other arg…

What is difference between UNION and UNION ALL in SQL Server

We use UNION and UNION ALL operator to combine multiple results set into one result set.
UNION operator is used to combining multiple results set into one result set but removes any duplicate rows. Basically, UNION is used to performing a DISTINCT operation across all columns in the result set. UNION operator has the extra overhead of removing duplicate rows and sorting result.
UNION ALL operator use to combine multiple results set into one result set but it does not remove any duplicate result. Actually, this does not remove duplicate rows so it is faster than the UNION operator. If you want to combine multiple results and without duplicate records then use UNION otherwise UNION ALL is better.
Following some rules for using UNION/UNION ALL operator
1.The number of the column should be the same in the query's when you want to combine them. 2.The column should be of the same data type. 3.ORDER BY clause can be applied to the overall result set not within each result set.
4.Column name of …