Wednesday, 21 August 2024

Select …. Into Vs Insert… select statement

To take the backup of the table either we can use Select …into or Insert …. Select statement. Here we will see which one is good to take the backup of the table.

Recently, we had taken a backup of the table. Table size is round 10GB. What we did first we have created a backup table with same structure and write script to load the data into the newly created backup table.

Insert into backup_table (List of columns)

Select list of column from table

To load this data it took more than 45 min and also we have observed that transaction log file grew.

Second time what we did instead of using the above approach we have taken the other approach to take the backup of the table.

Read: Select INTO in SQL Server

https://bageshkumarbagi-msbi.blogspot.com/2016/05/select-into-in-sql-server.html

It took less time to load the data into the backup table and also transaction log file not grew as compare to first approach.

Let’s see why it is select … into is faster than insert … select statement.

For the demo here we are using AdventureWorksDW2019 DB and FactProductInventory table. This table is having more than 776K records and size of this table is 30MB.


Before taking the backup checking the log size of this database.

SELECT file_id, name, type_desc,size,(size * 8)/1024.0 AS size_in_mb

FROM sys.database_files

Now creating table for the backup.

CREATE TABLE factproductinventory_bk_08_11_2024

  (

     productkey   INT NOT NULL,

     datekey      INT NOT NULL,

     movementdate DATE NOT NULL,

     unitcost     MONEY NOT NULL,

     unitsin      INT NOT NULL,

     unitsout     INT NOT NULL,

     unitsbalance INT NOT NULL

  )

Table created successfully 

We have enabled ‘STATISTICS TIME’ and ‘STATISTICS IO’ to view query (execution time and CPU) and disk I/O (read and writes) statistics. I also tracked the log space usage by looking at the ‘DBCC SQLPERF (LOGSPACE)’ view. 

Now loading the data.

As of now

select Operation,count(*) as [Number of Records] from fn_dblog(null,null)

where Operation in ('LOP_INSERT_ROWS','LOP_MODIFY_ROW') group by Operation order by 2 desc

 

Now inserting the records.

insert into factproductinventory_bk_08_11_2024

  (productkey,datekey,movementdate,unitcost,unitsin,unitsout,unitsbalance)

  select productkey,datekey,movementdate,unitcost,unitsin,unitsout,unitsbalance

  from factproductinventory

Data inserted successfully 

It took 4500 MS to insert the records. Logical read is 7780128.

See the log details


Log size is increase from 72.00 MB to 264 MB.

Log space use % also increased from 2.33% to 38.36%.

                            
See the 776K records are written into the transaction log.

See the records into the table 

Let’s the other approach

select * into factproductinventory_BK_08112024

  from factproductinventory

  

See here it took around 1200 MS to insert records into backup table and Logical read is 3860.

See the transaction log details.

 

Nothing written to the log. 

No changes on the log space used %.


This is one reason that it is faster.

Se the records in the backup table.


No comments:

Post a Comment

If you have any doubt, please let me know.

Popular Posts