次の方法で共有

TOP N Biggest Tables in every database

  • [アーティクル]

Sometime it may be useful to find all the big tables in all databases.

There may be performance issue on the system and it can be necessary to move some tables to another disk to decrease the disk latency. So additional File Groups and Files can be created and some of those big tables can be moved to those File Groups using create index or alter index command.

But How those big tables can be found easily?

The script below can be used to identify the big tables. It can be create as a stored procedure with parameters so database name or object name can be passed as a parameter.

The result can be saved in a history table and it will be good to see the increase rate for each table

 USE master;
GO
if not exists (select name from sys.databases where name = 'SQLAdmin')
    create database SQLAdmin
go
declare @dbname sysname
declare @dbid varchar(5)
declare @sql as varchar(4000)

CREATE TABLE #sys_table_sizes (
    database_id int,
    database_name SYSNAME, 
    schema_id int,
    schema_name sysname, 
    object_id int,
    object_name sysname, 
    row_count CHAR(11), 
    reserved_in_kb decimal(15,2),
    data_size_in_kb decimal(15,2),
    index_size_in_kb decimal(15,2),
    unused_size_in_kb decimal(15,2)
)

CREATE TABLE #sys_table_sizes_xml_and_fulltext (
    database_id int,
    schema_id int,
    object_id int,
    reserved_in_kb decimal(15,2),
    unused_size_in_kb decimal(15,2)
)

declare dbname_cursor CURSOR FOR
SELECT name, database_id from sys.databases 
where database_id not in (1,2,3,4) and state=0

OPEN dbname_cursor

FETCH NEXT FROM dbname_cursor
INTO @dbname, @dbid

WHILE @@FETCH_STATUS = 0
BEGIN

--in_row_data_page_count 
--If the partition is part of a heap, the value is the number of data pages in the heap. 
--If the partition is part of an index, the value is the number of pages in the leaf level. 
--Nonleaf pages in the B-tree are not included in the count.
--IAM (Index Allocation Map) pages are not included in either case.

--in_row_used_page_count 
--Total number of pages in use to store and manage the in-row data in this partition. 
--This count includes nonleaf B-tree pages, IAM pages, and all pages included in the in_row_data_page_count column.

--used_page_count
--Total number of pages used for the partition. 
--Computed as in_row_used_page_count + lob_used_page_count + row_overflow_used_page_count.

set @sql = '
insert into #sys_table_sizes
select ''' +
@dbid + ''' database_id, 
''' +
@dbname + ''' database_name, 
o.schema_id schema_id,
object_schema_name (ps.object_id, ' + @dbid + ') schema_name,
ps.object_id object_id,
object_name(ps.object_id, '+ @dbid + ') object_name, 
row_count = sum (case when (index_id<2) then row_count else ''0'' end),
(SUM(reserved_page_count)*8) as reserved_in_kb, 
data_size_in_kb = sum (case when (index_id<2) 
                            then (in_row_data_page_count + lob_used_page_count + row_overflow_used_page_count) 
                            else ''0'' end)*8,
index_size_in_kb = (sum(used_page_count) - sum (case when (index_id<2) 
                            then (in_row_data_page_count + lob_used_page_count + row_overflow_used_page_count) 
                            else ''0'' end))*8,
unused_size_in_kb = (SUM(reserved_page_count)-SUM(used_page_count))    *8                
from [' + @dbname + '].sys.dm_db_partition_stats ps
join [' + @dbname + '].sys.objects o on ps.object_id=o.object_id
where o.type = ''U'' and is_ms_shipped = 0
group by o.schema_id, ps.object_id
'

--print @sql
exec (@sql)

set @sql = '
insert into #sys_table_sizes_xml_and_fulltext
select ''' +
@dbid + ''' database_id, 
(select schema_id from [' + @dbname + '].sys.objects where object_id = it.parent_object_id) schema_id,
it.parent_object_id object_id,
(SUM(reserved_page_count)*8) as reserved_in_kb, 
unused_size_in_kb = (SUM(reserved_page_count)-SUM(used_page_count))    *8                
from [' + @dbname + '].sys.dm_db_partition_stats ps
join [' + @dbname + '].sys.objects o on ps.object_id=o.object_id
join [' + @dbname + '].sys.internal_tables it  on ps.object_id = it.object_id
where it.internal_type IN (202,204,211,212,213,214,215,216) 
group by o.schema_id, it.parent_object_id
'

--print @sql
exec (@sql)

   FETCH NEXT FROM dbname_cursor
   INTO @dbname, @dbid
END

update A
set A.reserved_in_kb = convert(int, A.reserved_in_kb) + convert(int, B.reserved_in_kb),
A.index_size_in_kb = convert(int, A.index_size_in_kb) + (convert(int, B.reserved_in_kb) - convert(int, B.unused_size_in_kb)),
A.unused_size_in_kb = convert(int, A.unused_size_in_kb) + convert(int, B.unused_size_in_kb)
from #sys_table_sizes A
join #sys_table_sizes_xml_and_fulltext B on A.database_id=B.database_id and A.schema_id = B.schema_id and A.object_id=B.object_id

CLOSE dbname_cursor
DEALLOCATE dbname_cursor

--select * from #sys_table_sizes 
--select * from #sys_table_sizes_xml_and_fulltext

/*
--more columns
select *, 
[index_size > data_size] = case when convert(int, index_size_in_kb)>convert(int, data_size_in_kb) then 'Yes' else 'No' end
from #sys_table_sizes 
order by database_name asc, reserved_in_kb desc
*/

/*
--all rows
select database_name, schema_name, object_name, row_count, reserved_in_kb, data_size_in_kb, index_size_in_kb, unused_size_in_kb, 
[index_size > data_size] = case when convert(int, index_size_in_kb)>convert(int, data_size_in_kb) then 'Yes' else 'No' end
from #sys_table_sizes 
order by database_name asc, reserved_in_kb desc
*/
go

--TOP N rows
--define a CTE with the name cte_table to get top 10 biggest tables in every database
;WITH cte_table AS (
     SELECT
        ROW_NUMBER() OVER ( PARTITION BY database_name ORDER BY reserved_in_kb DESC ) AS 'RowNumber',
        database_name, schema_name, object_name, row_count, reserved_in_kb, 
        data_size_in_kb, index_size_in_kb, unused_size_in_kb, 
        [index_size > data_size] = case when convert(int, index_size_in_kb)>convert(int, data_size_in_kb) 
                                        then 'Yes' else 'No' end
      FROM #sys_table_sizes
)
-- and select the data from the CTE table
SELECT
        database_name, schema_name, object_name, row_count, reserved_in_kb, 
        data_size_in_kb, index_size_in_kb, unused_size_in_kb, 
        [index_size > data_size] = case when convert(int, index_size_in_kb)>convert(int, data_size_in_kb) 
                                        then 'Yes' else 'No' end
FROM cte_table
WHERE RowNumber <= 10


/*
--Uncomment the lines for Testing
--Compare the result with sp_spaceused command
sp_spaceused 'Sales.Store'
go
select 'Sales.Store' as test_object,object_name, row_count, reserved_in_kb, data_size_in_kb, index_size_in_kb, unused_size_in_kb from #sys_table_sizes 
where database_name = 'AdventureWorks2008' and object_name like 'Store'
go
sp_spaceused 'Person.Person'
go
select 'Person.Person' as test_object, object_name, row_count, reserved_in_kb, data_size_in_kb, index_size_in_kb, unused_size_in_kb from #sys_table_sizes 
where database_name = 'AdventureWorks2008' and object_name like 'Person'
go
sp_spaceused 'Production.ProductModel' --xml data
go
select 'Production.ProductModel' as test_object, object_name, row_count, reserved_in_kb, data_size_in_kb, index_size_in_kb, unused_size_in_kb from #sys_table_sizes 
where database_name = 'AdventureWorks2008' and object_name like 'ProductModel'
go
*/
drop table #sys_table_sizes
drop table #sys_table_sizes_xml_and_fulltext

Comments