I need the run the following query on many databases, I have over 100+ databases, but I don't want to pull up each database and run the query one at a time.
The User table is only listed in Database#_Account.
If the query is ran it errors out because Database#_Admin does not have User Table.
(EXAMPLE Database List)
Database:
---------------------
MASTER
Model
msdb
tempdb
Database1_Account
Database1_Admin
Database2_Account
Database2_Admin
Database3_Account
Database3_Admin
Query:
EXEC sp_MsForEachDb #command1 = SELECT "?" as DatabaseName, *
FROM ?.User
WHERE Name = "John" AND "?" LIKE "%_Account"
ms_foreachDb is still an undocumented function and it subject to change anytime. I would use a cursor for something like this.
Here is a working template to get you started:
DECLARE #tsql nvarchar(max)
DECLARE #dbname varchar(500)
DECLARE MyCur CURSOR STATIC FORWARD_ONLY FOR
SELECT [name]
FROM sys.databases
WHERE [name] NOT IN ('tempdb')
OPEN MyCur
WHILE (1=1)
BEGIN
FETCH NEXT FROM MyCur INTO #dbname
IF ##FETCH_STATUS <> 0
BREAK
SET #tsql = 'use ' + #dbname + ' SELECT * FROM sys.tables'
EXEC sp_executesql #tsql
END
CLOSE MyCur;
DEALLOCATE MyCur;
You need to pass the command as an nvarchar literal, not as a query.
You need to use the correct nomenclature. You've left out the schema name. It's Database.Schema.Table, not Database.Table. I'm assuming all tables use the default dbo schema.
Write the query to test if the table exists before executing. Easiest way to do that is with IF OBJECT_ID(N'TableName') IS NOT NULL.
Avoid double quotes. They're normally field identifiers like square brackets are, so they're potentially ambiguous when used for varchar literals.
Try:
EXEC sp_MsForEachDb #command1 = N'IF OBJECT_ID(N''?.dbo.User'') IS NOT NULL SELECT ''?'' as DatabaseName, * FROM ?.dbo.User WHERE Name = ''John'' AND ''?'' LIKE ''%_Account'''
Here's the query I use to do a while loop to iterate through Databases.
Just put your code where it says PUT CODE HERE.
SET NOCOUNT ON
DECLARE #Database TABLE (DbName SYSNAME)
DECLARE #DbName AS SYSNAME
SET #DbName = ''
INSERT INTO #Database (DbName)
SELECT NAME
FROM master.dbo.sysdatabases
WHERE NAME <> 'tempdb'
ORDER BY NAME ASC
WHILE #DbName IS NOT NULL
BEGIN
SET #DbName = (
SELECT MIN(DbName)
FROM #Database
WHERE DbName > #DbName
)
/*
PUT CODE HERE
EXAMPLE PRINT Database Name
*/
PRINT #DbName
END
To create a list of users that match certain conditions you can modify this script.
Is there a query which drops / deletes databases with no tables in them (deletes empty databases)?
Server is Microsoft SQL Server 2005
This should do it.
Tested on a lab machine and it dropped all databases with 0 user tables.
Note, however, that tables aren't the only things in a database, necessarily. There could be stored procedures, functions, etc that someone might still need.
NOTE THAT THIS IS A VERY DANGEROUS OPERATION, AS IT DROPS DATABASES. USE AT YOUR OWN RISK. I AM NOT RESPONSIBLE FOR DAMAGE YOU CAUSE.
USE [master];
DECLARE #name varchar(50);
DECLARE #innerQuery varchar(max);
DECLARE tableCursor CURSOR FOR SELECT name FROM sys.databases where owner_sid != 0x01;
OPEN tableCursor;
FETCH NEXT FROM tableCursor
INTO #name
WHILE ##FETCH_STATUS = 0
BEGIN
SET #innerQuery = 'USE [' + #name + ']; IF (SELECT COUNT(*) FROM sys.objects WHERE type = ''U'') = 0
BEGIN
USE [master];
DROP DATABASE [' + #name + ']
END'
EXEC(#innerQuery)
FETCH NEXT FROM tableCursor INTO #name
END
CLOSE tableCursor;
DEALLOCATE tableCursor;
Note also that, if a database is in use, SQL Server will refuse to drop it. So, if there are other connections to a particular database that this tries to drop, the command will abort.
To avoid that problem, you can set the database in question to single-user mode.
The following script is the same as the above, except it also sets the target databases to single-user mode to kill active connections.
BE EVEN MORE CAREFUL WITH THIS, AS IT'S ESSENTIALLY THE NUCLEAR OPTION:
use [master];
DECLARE #name varchar(50);
DECLARE #innerQuery varchar(max);
DECLARE tableCursor CURSOR FOR SELECT name FROM sys.databases where owner_sid != 0x01;
OPEN tableCursor;
FETCH NEXT FROM tableCursor
INTO #name
WHILE ##FETCH_STATUS = 0
BEGIN
SET #innerQuery =
'USE [' + #name + '];
IF (SELECT COUNT(*) FROM sys.objects WHERE type = ''U'') = 0
BEGIN
USE [master];
ALTER DATABASE [' + #name + '] SET SINGLE_USER WITH ROLLBACK IMMEDIATE;
DROP DATABASE [' + #name + '];
END'
EXEC(#innerQuery)
FETCH NEXT FROM tableCursor INTO #name
END
CLOSE tableCursor;
DEALLOCATE tableCursor;
I think it is not possible (at least not with a TSQL command, might be a stored procedure somewhere). You would have to query sysobjects and abort if any found. (And you have to decide if you want to ignore some of the system objects like the design tables).
The script below will generate the needed DROP DATABASE script. You could tweak this to execute the statement too (in the master database context) but I suggest you review it first just in case.
EXEC sp_MSforeachdb N'
USE [?];
IF N''?'' NOT IN(N''master'', N''model'', N''msdb'', N''tempdb'')
BEGIN
IF NOT EXISTS(
SELECT *
FROM sys.tables
WHERE
OBJECTPROPERTYEX(object_id, ''IsMSShipped'') = 0
)
BEGIN
PRINT ''DROP DATABASE [?];'';
END;
END;';
We have a few test databases that we are throwing test indexes into and the log file gets bloated real quickly due to us dropping the contents of the table and repopulating it.
I have found, thanks to Stack Overflow, a few scripts and put them together to do what I need.
Here is the script:
USE SSSIndexes
GO
ALTER DATABASE SSSIndexes SET RECOVERY SIMPLE WITH NO_WAIT
GO
DBCC SHRINKFILE(N'SSSIndexes_Log', 1) <-- my issue is here
GO
The problem is the log file name. Is there a way to get the log file name without having to look it up manually and include it in the script to where this part is automated?
Btw, we never intend on restore this database. These are temporary indexes.
Thanks!
USE SSSIndexes
GO
ALTER DATABASE SSSIndexes SET RECOVERY SIMPLE WITH NO_WAIT
GO
DECLARE #Name NVARCHAR(50)
DECLARE cur CURSOR FOR
SELECT [name]
FROM [sys].[database_files]
where [type] = 1
OPEN cur
FETCH NEXT FROM cur INTO #Name
WHILE ##FETCH_STATUS = 0
BEGIN
DBCC SHRINKFILE(#Name, 1)
FETCH NEXT FROM cur INTO #Name
END
CLOSE cur
DEALLOCATE cur
You can use this to generate script for log file truncating of all databases on a specific server. To stick to specific databases use a filter.
SELECT
' USE [' + name + ']; GO
-- Truncate the log by changing the database recovery model to SIMPLE.
ALTER DATABASE [' + name + '] SET RECOVERY SIMPLE;
GO
-- Shrink the truncated log file to 1 MB.
DECLARE #logname varchar(128);
SELECT TOP 1 #logname=[name] FROM [' + name + '].[sys].[database_files] WHERE [type] = 1;
DBCC SHRINKFILE (#logname, 1);
GO
-- Reset the database recovery model.
ALTER DATABASE [' + name + '] SET RECOVERY FULL;
GO
' AS qry FROM master.dbo.sysdatabases
WHERE dbid > 6
WARNING!!!: YOU SHOULD ONLY DO THIS ON TEST DB SERVERS. Production DBs typically WANT to have FULL recovery mode. Test DBs you usually don't care.
Expanding on Abdul's answer (I had trouble getting it to show in new lines) and Dis' answers this finds the first db that has recovery FULL, sets it to simple and shrinks it... just keep clicking execute until it returns null
declare #dbname nvarchar(50)
declare #logfilename nvarchar(100)
select top(1) #dbname = name from sys.databases where recovery_model_desc <> 'SIMPLE' AND name <> 'tempdb'
declare #shrinkdb nvarchar(500)
set #shrinkdb = 'USE master
ALTER DATABASE [' + #dbname + '] SET RECOVERY SIMPLE
'
select #shrinkdb
execute sp_executesql #shrinkdb
set #shrinkdb = 'USE [' + #dbname + ']
DECLARE #Name NVARCHAR(50)
DECLARE cur CURSOR FOR
SELECT [name]
FROM [sys].[database_files]
where [type] = 1
OPEN cur
FETCH NEXT FROM cur INTO #Name
WHILE ##FETCH_STATUS = 0
BEGIN
DBCC SHRINKFILE(#Name, 1)
FETCH NEXT FROM cur INTO #Name
END
CLOSE cur
DEALLOCATE cur
'
select #shrinkdb
execute sp_executesql #shrinkdb
I have a development database that re-deploy frequently from a Visual Studio Database project (via a TFS Auto Build).
Sometimes when I run my build I get this error:
ALTER DATABASE failed because a lock could not be placed on database 'MyDB'. Try again later.
ALTER DATABASE statement failed.
Cannot drop database "MyDB" because it is currently in use.
I tried this:
ALTER DATABASE MyDB SET RESTRICTED_USER WITH ROLLBACK IMMEDIATE
but I still cannot drop the database. (My guess is that most of the developers have dbo access.)
I can manually run SP_WHO and start killing connections, but I need an automatic way to do this in the auto build. (Though this time my connection is the only one on the db I am trying to drop.)
Is there a script that can drop my database regardless of who is connected?
Updated
For MS SQL Server 2012 and above
USE [master];
DECLARE #kill varchar(8000) = '';
SELECT #kill = #kill + 'kill ' + CONVERT(varchar(5), session_id) + ';'
FROM sys.dm_exec_sessions
WHERE database_id = db_id('MyDB')
EXEC(#kill);
For MS SQL Server 2000, 2005, 2008
USE master;
DECLARE #kill varchar(8000); SET #kill = '';
SELECT #kill = #kill + 'kill ' + CONVERT(varchar(5), spid) + ';'
FROM master..sysprocesses
WHERE dbid = db_id('MyDB')
EXEC(#kill);
USE master
GO
ALTER DATABASE database_name
SET OFFLINE WITH ROLLBACK IMMEDIATE
GO
Ref: http://msdn.microsoft.com/en-us/library/bb522682%28v=sql.105%29.aspx
You can get the script that SSMS provides by doing the following:
Right-click on a database in SSMS and choose delete
In the dialog, check the checkbox for "Close existing connections."
Click the Script button at the top of the dialog.
The script will look something like this:
USE [master]
GO
ALTER DATABASE [YourDatabaseName] SET SINGLE_USER WITH ROLLBACK IMMEDIATE
GO
USE [master]
GO
DROP DATABASE [YourDatabaseName]
GO
Little known: the GO sql statement can take an integer for the number of times to repeat previous command.
So if you:
ALTER DATABASE [DATABASENAME] SET SINGLE_USER
GO
Then:
USE [DATABASENAME]
GO 2000
This will repeat the USE command 2000 times, force deadlock on all other connections, and take ownership of the single connection. (Giving your query window sole access to do as you wish.)
To my experience, using SINGLE_USER helps most of the times, however, one should be careful: I have experienced occasions in which between the time I start the SINGLE_USER command and the time it is finished... apparently another 'user' had gotten the SINGLE_USER access, not me. If that happens, you're in for a tough job trying to get the access to the database back (in my case, it was a specific service running for a software with SQL databases that got hold of the SINGLE_USER access before I did).
What I think should be the most reliable way (can't vouch for it, but it is what I will test in the days to come), is actually:
- stop services that may interfere with your access (if there are any)
- use the 'kill' script above to close all connections
- set the database to single_user immediately after that
- then do the restore
Matthew's supremely efficient script updated to use the dm_exec_sessions DMV, replacing the deprecated sysprocesses system table:
USE [master];
GO
DECLARE #Kill VARCHAR(8000) = '';
SELECT
#Kill = #Kill + 'kill ' + CONVERT(VARCHAR(5), session_id) + ';'
FROM
sys.dm_exec_sessions
WHERE
database_id = DB_ID('<YourDB>');
EXEC sys.sp_executesql #Kill;
Alternative using WHILE loop (if you want to process any other operations per execution):
USE [master];
GO
DECLARE #DatabaseID SMALLINT = DB_ID(N'<YourDB>');
DECLARE #SQL NVARCHAR(10);
WHILE EXISTS ( SELECT
1
FROM
sys.dm_exec_sessions
WHERE
database_id = #DatabaseID )
BEGIN;
SET #SQL = (
SELECT TOP 1
N'kill ' + CAST(session_id AS NVARCHAR(5)) + ';'
FROM
sys.dm_exec_sessions
WHERE
database_id = #DatabaseID
);
EXEC sys.sp_executesql #SQL;
END;
The accepted answer has the drawback that it doesn't take into consideration that a database can be locked by a connection that is executing a query that involves tables in a database other than the one connected to.
This can be the case if the server instance has more than one database and the query directly or indirectly (for example through synonyms) use tables in more than one database etc.
I therefore find that it sometimes is better to use syslockinfo to find the connections to kill.
My suggestion would therefore be to use the below variation of the accepted answer from AlexK:
USE [master];
DECLARE #kill varchar(8000) = '';
SELECT #kill = #kill + 'kill ' + CONVERT(varchar(5), req_spid) + ';'
FROM master.dbo.syslockinfo
WHERE rsc_type = 2
AND rsc_dbid = db_id('MyDB')
EXEC(#kill);
You should be careful about exceptions during killing processes. So you may use this script:
USE master;
GO
DECLARE #kill varchar(max) = '';
SELECT #kill = #kill + 'BEGIN TRY KILL ' + CONVERT(varchar(5), spid) + ';' + ' END TRY BEGIN CATCH END CATCH ;' FROM master..sysprocesses
EXEC (#kill)
#AlexK wrote a great answer. I just want to add my two cents. The code below is entirely based on #AlexK's answer, the difference is that you can specify the user and a time since the last batch was executed (note that the code uses sys.dm_exec_sessions instead of master..sysprocess):
DECLARE #kill varchar(8000);
set #kill =''
select #kill = #kill + 'kill ' + CONVERT(varchar(5), session_id) + ';' from sys.dm_exec_sessions
where login_name = 'usrDBTest'
and datediff(hh,login_time,getdate()) > 1
--and session_id in (311,266)
exec(#kill)
In this example only the process of the user usrDBTest which the last batch was executed more than 1 hour ago will be killed.
You can use Cursor like that:
USE master
GO
DECLARE #SQL AS VARCHAR(255)
DECLARE #SPID AS SMALLINT
DECLARE #Database AS VARCHAR(500)
SET #Database = 'AdventureWorks2016CTP3'
DECLARE Murderer CURSOR FOR
SELECT spid FROM sys.sysprocesses WHERE DB_NAME(dbid) = #Database
OPEN Murderer
FETCH NEXT FROM Murderer INTO #SPID
WHILE ##FETCH_STATUS = 0
BEGIN
SET #SQL = 'Kill ' + CAST(#SPID AS VARCHAR(10)) + ';'
EXEC (#SQL)
PRINT ' Process ' + CAST(#SPID AS VARCHAR(10)) +' has been killed'
FETCH NEXT FROM Murderer INTO #SPID
END
CLOSE Murderer
DEALLOCATE Murderer
I wrote about that in my blog here:
http://www.pigeonsql.com/single-post/2016/12/13/Kill-all-connections-on-DB-by-Cursor
SELECT
spid,
sp.[status],
loginame [Login],
hostname,
blocked BlkBy,
sd.name DBName,
cmd Command,
cpu CPUTime,
memusage Memory,
physical_io DiskIO,
lastwaittype LastWaitType,
[program_name] ProgramName,
last_batch LastBatch,
login_time LoginTime,
'kill ' + CAST(spid as varchar(10)) as 'Kill Command'
FROM master.dbo.sysprocesses sp
JOIN master.dbo.sysdatabases sd ON sp.dbid = sd.dbid
WHERE sd.name NOT IN ('master', 'model', 'msdb')
--AND sd.name = 'db_name'
--AND hostname like 'hostname1%'
--AND loginame like 'username1%'
ORDER BY spid
/* If a service connects continously. You can automatically execute kill process then run your script:
DECLARE #sqlcommand nvarchar (500)
SELECT #sqlcommand = 'kill ' + CAST(spid as varchar(10))
FROM master.dbo.sysprocesses sp
JOIN master.dbo.sysdatabases sd ON sp.dbid = sd.dbid
WHERE sd.name NOT IN ('master', 'model', 'msdb')
--AND sd.name = 'db_name'
--AND hostname like 'hostname1%'
--AND loginame like 'username1%'
--SELECT #sqlcommand
EXEC sp_executesql #sqlcommand
*/
USE MASTER
GO
DECLARE #Spid INT
DECLARE #ExecSQL VARCHAR(255)
DECLARE KillCursor CURSOR LOCAL STATIC READ_ONLY FORWARD_ONLY
FOR
SELECT DISTINCT SPID
FROM MASTER..SysProcesses
WHERE DBID = DB_ID('dbname')
OPEN KillCursor
-- Grab the first SPID
FETCH NEXT
FROM KillCursor
INTO #Spid
WHILE ##FETCH_STATUS = 0
BEGIN
SET #ExecSQL = 'KILL ' + CAST(#Spid AS VARCHAR(50))
EXEC (#ExecSQL)
-- Pull the next SPID
FETCH NEXT
FROM KillCursor
INTO #Spid
END
CLOSE KillCursor
DEALLOCATE KillCursor
If you want to only drop/delete a database you can select the option "Close existing connections" which is by default unset.
Right-click on Database catalog -> options.
Delete -> check option.
Ok
I have tested successfully with simple code below
USE [master]
GO
ALTER DATABASE [YourDatabaseName] SET SINGLE_USER WITH ROLLBACK IMMEDIATE
GO
So, I have a database with around 100 audit tables and I want to empty them out preferably in 1 sql query. The database has 2 sets of tables [audit] and [dbo]. I'm not even sure what I should be calling these groups of tables so finding any kind of results from google is proving difficult.
Any suggestions?
You can find all tables with a certain schema name like:
select name from sys.tables where schema_name(schema_id) = 'audit'
With a cursor, you iterate over those tables, and empty them using TRUNCATE TABLE:
use db
declare #query nvarchar(max)
declare #tablename nvarchar(max)
declare #curs cursor
set #curs = cursor for select name from sys.tables
where schema_name(schema_id) = 'audit'
open #curs
fetch next from #curs into #tablename
while ##FETCH_STATUS = 0
begin
set #query = N'truncate table audit.' + #tablename
exec sp_executesql #query
fetch next from #curs into #tablename
end
close #curs
deallocate #curs
If you want to delete the tables instead, use:
set #query = N'drop table audit.' + #tablename
You could also use the sp_msforeachtable stored procedure.
It lets you perform a query on each user table in the current DB.
For example the following will truncate all user tables in your DB
Use YourDB
Exec sp_msforeachtable 'TRUNCATE TABLE ?'
And this will truncate all user tables in the specified db that belong to the audit schema.
Use YourDB
Exec sp_msforeachtable #command1 = '
if (Select Object_Schema_name(object_id(''?''))) = ''dbo''
Begin
TRUNCATE TABLE ?
print ''truncated '' + ''?''
End
'
Also Here is a blog entry with some more uses for this stored procedure.