Using Triggers To Track Changes In A Single Audit Table
Feb 6, 2006
Hi,
I am looking to track any changes made to any table within a db into a single audit table which will hold as fields: the table that has been updated/inserted, the field that was changed, its primary key, the old value and the new value specific to that field, and the date it was updated/inserted.
From what I have read, it does not look like this is possible with a trigger on table as it is not row specific and that I might have to control this from the business layer (vb.net). I am correct in this assumption, or is there a way of tracking specific data changes through triggers.
Im using triggers to track changes Insert/Update/Deletes on my DB tables and they work for when i am manually adding/editing and deleting a single records.
The problem arises in that I have an asset/inventory management app that dumps lots of details into my DB tables at once each time its run. Not all of the tables are updated and data cannot be completely inserted.
This is the trigger i have been using - could someone tell me how to modify it to work.
/* This trigger audit trails all changes made to a table. It will place in the table Audit all inserted, deleted, changed columns in the table on which it is placed. It will put out an error message if there is no primary key on the table You will need to change @TableName to match the table to be audit trailed */
ALTER trigger tr_TableName on dbo.TableName for insert, update, delete as
declare @bit int , @field int , @maxfield int , @char int , @fieldname varchar(128) , @TableName varchar(128) , @PKCols varchar(1000) , @sql varchar(2000), @UpdateDate varchar(21) , @Action nvarchar(50) , @HostName nvarchar(50), @PKFieldName varchar (1000)
IF EXISTS(SELECT * FROM inserted) IF EXISTS(SELECT * FROM deleted) --update = inserted and deleted tables both contain data BEGIN SET @Action = 'UPDATE' SELECT @DeviceID = (SELECT inserted.DeviceID FROM inserted INNER JOIN deleted ON inserted.deviceID = deleted.deviceid) END ELSE
--insert = inserted contains data, deleted does not BEGIN SET @Action = 'INSERT' select @DeviceID = (SELECT DeviceID from inserted) END ELSE --delete = deleted contains data, inserted does not BEGIN SET @Action = 'DELETE' select @DeviceID = (SELECT DeviceID from deleted) END
select @TableName = 'TableName'
-- date select @HostName = host_name(), @UpdateDate = convert(varchar(8), getdate(), 112) + ' ' + convert(varchar(12), getdate(), 114), --@DeviceID, @PKFieldName=(select top 1 c.COLUMN_NAME fromINFORMATION_SCHEMA.TABLE_CONSTRAINTS pk ,INFORMATION_SCHEMA.KEY_COLUMN_USAGE c where pk.TABLE_NAME = @TableName andCONSTRAINT_TYPE = 'PRIMARY KEY'andc.TABLE_NAME = pk.TABLE_NAME and c.CONSTRAINT_NAME = pk.CONSTRAINT_NAME)
-- get list of columns select * into #ins from inserted select * into #del from deleted
-- Get primary key columns for full outer join select@PKCols = coalesce(@PKCols + ' and', ' on') + ' i.' + c.COLUMN_NAME + ' = d.' + c.COLUMN_NAME fromINFORMATION_SCHEMA.TABLE_CONSTRAINTS pk , INFORMATION_SCHEMA.KEY_COLUMN_USAGE c where pk.TABLE_NAME = @TableName andCONSTRAINT_TYPE = 'PRIMARY KEY' andc.TABLE_NAME = pk.TABLE_NAME andc.CONSTRAINT_NAME = pk.CONSTRAINT_NAME
if @PKCols is null begin raiserror('no PK on table %s', 16, -1, @TableName) return end
Hello.I tried to implement audit trail, by making an audit trail table with thefollowing fileds:TableName,FieldName,OldValue,NewValue,UpdateDate,t ype,UserName.Triggers on each table were set to do the job and everything was fine exceptthat in the audit trail you couldn't know which row exacltly wasupdated/inserted/deleted...Therefore I introduced 3 additional columnes(RowMark1, RowMark2, RowMark3) which should identify theinserted/updated/deleted row.For example, RowMark1 could be foreign key, RowMark2 could be primary key,and RowMark3 could be autonumber ID.But, when I have several rows updated, RowMark columnes values are identicalin all rows in the audit trail table! What is wrong with my code, and how tosolve it ?Thank you in advance!CREATE TRIGGER Trigger_audit_TableNameON dbo.TableNameFOR DELETE, INSERT, UPDATEAS BEGINdeclare @type nvarchar(20) ,@UpdateDate datetime ,@UserName nvarchar(100),@RowMark1 nvarchar (100),@RowMark2 nvarchar (100),@RowMark3 nvarchar (100)if exists (select * from inserted) and exists (select * fromdeleted)select @type = 'UPDATE',@RowMark1=d.ForeignKeyField,@RowMark2=d.PrimaryKey Field,@RowMark3=d.IDfrom deleted delse if exists (select * from inserted)select @type = 'INSERT',@RowMark1=i.ForeignKeyField,@RowMark2=i.PrimaryKey Field,@RowMark3=i.IDfrom inserted ielseselect @type = 'DELETE',@RowMark1=d.ForeignKeyField,@RowMark2=d.PrimaryKey Field,@RowMark3=d.IDfrom deleted dselect @UpdateDate = getdate() ,@UserName = USER/*The following code is repeated for every field in a table*/if update (FieldName) or @type = 'DELETE'insert dbo.AUDIT_TRAIL (TableName, FieldName, OldValue, NewValue,UpdateDate, UserName, type,RowMark1,RowMark2,RowMark3)select 'Descriptive Table Name', convert(nvarchar(100), 'DescriptiveField Name'),convert(nvarchar(1000),d.FieldName),convert(nvarchar(1000),i.FieldName),@UpdateDate, @UserName, @type, @RowMark1, @RowMark2,@RowMark3from inserted ifull outer join deleted don i.ID = d.IDwhere (i.FieldName <> d.FieldNameor (i.FieldName is null and d.FieldName is not null)or (i.FieldName is not null and d.FieldName is null))END
Dear Group,I would like to create an audit table that is created with a trigger thatreflects all the changes(insert, update and delete) that occur in table.Say I have a table withSubject_ID, visit_number, dob, weight, height, User_name, inputdateThe audit table would have .Subject_ID, visit_number, dob, weight, height, User_name, inputdate,edit_action, edit_reason.Where the edit_action would be insert, update, delete; the edit_reason wouldbe the reason given for the edit.Help with this would be great, since I am new to the world of triggers.Thanks,Jeff
HiI am looking to implement an audit/history table/tables but am lookingat doing this without the use of triggers.The reason for doing this is that the application is highlytransactional and speed in critical areas is important.I am worried that triggers would slow things down.I am more used to other database where by there is a utility to "dump"the contents of the transaction logs and use this for auditingpurposes. However SQL Server does not have this functionality (unlessthere is a sql server tool - 3rd party that I do not know about)Has anyone implemented something similar? Or used/using a 3rd partytool that will do this job.Effectively the clients would like to "look" at what happened - say 15minutes ago.thanksjohn
We have a particular database sat on SQL Server 2012 box along with about 20 other databases.
What I require is a method/Script/Audit that will simply track anyone who logs (successfully / unsuccessful) into this one particular database on the server (The single database is the key as the end user does not want information on any of the other databases that sit on the server), it also has to log time the attempt was made and it must track the logins via SQL Server or the application itself that is attached to the database.
I have ddl triggers in place to watch what people do to our various database environments. I can see when someone does something to a login, but I can't tell what was done. I have a sneaky someone creating accounts with sysadmin privs and I want to catch the source. I also want to know when someone changes a password on a sql account. Does anyone know of a way to do this?
i'm in a bit of a bind at work. if anyone could help, i'd greatlyappreciate it.i have a web app connecting to a sql server using sql serverauthentication. let's say, for example, my login/password isdbUser/dbUser. the web app however, is using windows authentication.so if I am logged into the network as 'DOMAINEric', when I access myweb app, my web app knows that I am 'DOMAINEric'. but to the sqlserver db, I am user 'dbUser'.now, i for each table i have, i need to implement an audit table torecord all updates, inserts, deletes that occur against it. i wasgoing to do so with triggers. this is all fine for selects, inserts,and updates. for each table, i have an updatedby and an updatedate.for example, let's say i have a table:create table blah(id int,col1 varchar(10),updatedby varchar(30),updatedate datetime)and corresponding audit table:create audit_blah(id int,blah_id int,blah_col1 varchar(10),blah_updatedby varchar(1),blah_updatedate datetime)for update and insert triggers, i can know what to insert into theupdatedby column of audit_blah because it's in a corresponding row inblah. my web app knows what user is accessing the application, andcan insert that name into blah. blah's trigger will then insert thatname into audit_blah.however, in the case of a delete, i'm not passing in an 'updatedby',because i'm deleting. in this situation, how can the trigger knowwhat user is deleting? the db only knows that sql user 'dbUser' isdeleting, but doesn't know that 'dbUser' is deleting on behalf of'DOMAINEric'. is there any way for my app to inform the trigger toaccess my windows identity without having a corresponding row in thetable from which to pull that info?obviously, i could have each of my app's users log into SQL serverthrough Windows authentication; then i could just use SYSTEM_USER.but let's say, for performance's sake, it'd be better for me to useone sql server login. (i believe one user works better for connectionpooling purposes.) is there a way to get around this?(i'm hoping a built-in function exists that solves all my problems.)suggestions? resources?any help would be great appreciated.happy turkeys.Eric
Hi, I am trying to figure out which of these option is best suited for auditing. Although each one of them has its own pros/cons. CLR trigger is easy to write and can be made generic so that it fits for any table required to be audited. I tried both the option in test database and i found the CLR trigger performed poorly. Results were : For table A (3 columns) with TSQL trigger took less than a sec for 2500 sequential inserts. While table B (3 columns) having same structure with CLR trigger took more than 20 sec for 2500 sequential inserts.
Has anybody done performance comparision of this 2 approaches ? Please share results if any.
I wanted to validate that is my findings correct so that i select best optimized approach.
I have made a server security audit and specify from database audit specification to audit "select" on a certain user and on a certain table. I logged in by this user and made the select statement..when i run this query
"select * from sys.fn_get_audit_file('d:Auditaudit1*',null,null)"
It return a value at which time the query has done
after 15 minutes i repeated the same action, i run the audit query and the same result is showed off on the panel.is it suppose to return a list of values by how many times this user has made the select statement on that table ? for example at 5:00 pm then 6:00 pm and so on
Hello. Taking a typical use having a history table, maintained from a base table via triggers... Trying to see how/if that can be done using the SQl 2005 Service Broker method, with messaging? The thought is that if we can do the History table updates ASYNC, the user will not wait more than setting up the Broker message queue. I saw this article about something similar, but it deals with LOGON triggers.
I'd think you can't do Hisotyr type triggers, with a message, because wouldn't you need to write all teh INSERTED/UPDATED data somewhere anyways? and there could be multiple rows affected in any given insert/update/delete, so could you even pass that thru to a Broker?
Anyone know of any references to using Broker Services for sending INSERTED/UPDATED data along for Historical versioning?
Also, was curious about error handling, because say you update teh base table, and then a problem occurs, and the Hisotry table is not updated. I want them in sync. Where is the message data stored, and is it accesible even if teh server reboots before the data is RECEIVED from teh QUEUE?
I am writing an adapter for a two way integration between two databases. The table I am currently working on has an update and insert trigger to send data to the other database via service broker.
Asd I insert/update data via my adapter I don't want the triggers to fire but I do want them to fire if somebody else make data modifications while I am working with the table.
I am doing all mu modifications within a transaction but I found I couldn't disable the trigger within a stored procedure to keep the disable and enable as close together as possible.
Does anybody ahve any ideas how to overcome this scenario or am I worrying about nothing as my transaction will keep other users from modifying the data until I've finished.
This isn€™t an problem as such, it€™s more of a debate.
If a table needs a number of update triggers which do differing tasks, should these triggers be separated out or encapsulated into one all encompassing trigger. Speaking in terms of performance, it doesn€™t make much of an improvement doing either depending upon the tasks performed. I was wondering in terms of maintenance and best practice etc. My view is that if the triggers do totally differing tasks they should be a trigger each on their own.
Hi, we are almost finished developing our database and we have a table we want to monitor because it is getting information deleted from it and it has a delete trigger on it but we want to track the changes to the table and were wondering how to track specific changes to a user database? We want to see who is making the change, what the change they are making is and also what is the time they are making it. I have used SELECT * FROM SYSPROCESSES and I am running SQL TRACE with filter on MS SQLEW, and MS TRANS, and Visual Basic with SQL statements on tblRoute,( the table that I want to monitor) and I want to know if there is any other way to monitor this table more closely?
Hi anyone please help! I have created the database driven web application with asp.net and sql server 2000. now I want keep track three operation(insert, updata and delete) that have been made on tables in a SQL Server 2000 database. what i did is: 1, create a audit table with columns: auditTable, actions, actionUser, actionTime 2, create three trigger(insert, update and delect respectivily) for every table my problem is that i can not get right user name. I use form authentication and i stored user login information in the database. every time, no matter who is logining to the web application, the action user is always SA. I user user-name() function to get userName(actionUser). Please anyone can help me to get current login user name, or tell the best way to track operations on a table. Thanks jili
I'm building an application that tracks courses and equivalent courses. For example, say I have course A which a user should get credit for if they have taken equivalent courses. So if a user has taken (course B or course C) AND (course D or course E) he should get credit for Course A. What's the best way to setup a SQL table to create these type of equivalent relationships?
Working on a "social networking site" and would like to have something that records a member's interactions within the site.So for instance if a member uploads a photo, I would like to record it and then display it as "[Member] added photo" with a link to the photo. Also, if a friend is accepted as a friend "[Member1] is now friends with [Member2]" and also if a member posts a comment on the forums "[Member] posted message in [ForumName]" with link to post.This is very similar to Facebook's News Feeds and MySpace Friend Updates.The only way I can think of having this is having an Activity table with a field for each different Id I want to record and an associated list of tracked Actions:Activity:idmemberIdactionTypePhotoIdMember2IdForumIdThen there would be the ActionType tableactionIdNameThe biggest problem I see with this solution is that I would need to add a new field in the Activity table for each different type of activity I would like to track. This could get pretty big as I begin to track more and more activity items.Is there some way I can generalize this. - Andy
Hi all, please help. I m trying to create an "empty" table from existing table for the audit trigger purpose. For now, i am trying to create an empty audit table for every table in a database named "pubs", and it's seem won't work. Please advise.. Thanks in advance.
SELECT @TABLE_NAME= MIN(TABLE_NAME) FROM INFORMATION_SCHEMA.TABLES WHERE TABLE_TYPE= 'BASE TABLE' AND TABLE_NAME NOT LIKE 'audit%' AND TABLE_NAME!= 'sysdiagrams' AND TABLE_NAME!= 'Audit' AND TABLE_NAME = 'sales'
WHILE @TABLE_NAME IS NOT NULL BEGIN
SELECT @TABLE_NAME= MIN(TABLE_NAME) FROM INFORMATION_SCHEMA.Tables WHERE TABLE_NAME> @TABLE_NAME AND TABLE_NAME = 'sales'
SELECT @TABLE_NAME= MIN(TABLE_NAME) FROM INFORMATION_SCHEMA.Tables WHERE TABLE_NAME> @TABLE_NAME AND TABLE_TYPE= 'BASE TABLE' AND TABLE_NAME!= 'sysdiagrams' AND TABLE_NAME!= 'Audit' AND TABLE_NAME NOT LIKE 'audit%'
I have a specific requirement. I need to insert the DML statements executed from Management Studio into a SQL table. We have SQL Server 2008 R2 and 2012 instances.
I would like to know about the DML process (Insert/update /delete) in a particular table .. it is like change tracking but I also want to know the modification date
I know CDC ( Change data capture ) but unfortunately it needs SQL 2008 developer/enterprise edition and my SQL server is SQL 2008 STANDARD edition.
I write a insert trigger on my table LeaveRegister(1000 rows) and inserting rows in audit table, but when i inserting a row in LeaveRegister table. In audit table 1000 + 1 rows are inserting every time.
Hallo i have two tables, MainTable and MainTableAudit: the second one keeps DML auditing via triggers. I'm trying to build a query to highlight changes occurred to fields between the MainTable record and its audited records in MainTableAudit, for example, let's suppose i entered an item with some wrong attributes, and edited it three times: MainTable record contains the latest and current version ID002|Hitchhikers Guide to the Galaxy|Sci-fi|240 pages MainTableAudit contains edited ID002 versions ID002|Hitchhikers Guide to the Galaxy|Sci-fi|232 pages|2006-07-08 08:32:12 ID002|Hitchhikers Guide to the Galaxy|Sci-fi|212 pages|2006-05-08 10:54:02 ID002|Hitchhikers Guide to Galaxy|Sci-fi|222 pages|2006-07-04 11:42:16
I would like to build a report like this: first insertion: Hitchhikers Guide to Galaxy|Sci-fi|222 pages modified on 2006-07-04 11:42:16: field "Title" changed from "Hitchhikers Guide to Galaxy" to "Hitchhikers Guide to the Galaxy", field "PageNo" changed from "222" to "212" modified on 2006-05-08 10:54:02: field "PageNo" changed from "212" to "232" modified on 2006-07-08 08:32:12: field "PageNo" changed from "232" to "240" current version: Hitchhikers Guide to the Galaxy|Sci-fi|240 pages
i'd prefer to use T-SQL and keep all into a single place (a view or storedprocedure), or at least to use reporting services; btw i would like to avoid coding web pages or hosted applications.
I need help...here is the problem.Last weekend, the servers in our datacenter where moved around. After thismove, and maybe coincidental, 1 server is performing very poor. Afterrunning a trace with SQL Profiler, I saw the problem which was laterconfirmed with another tool for SQL server performance monitoring. It seemsthat all connections to the SQL server (between 200 - 400) are doing a login/ logout for each command that they process. For example, the user'sconnection will login, perform a SELECT, and then logout. This is not a..NET application. The client software was not changed, it is still thesame. The vendor has said that it is not supposed to do that, it issupposed to use 1 connection that log's on in the morning and logs off atthe end of the day or whenever the user exits. 1 user may have severalconnections to the database.At times, the server is processing over 250 login / logouts (avgeraged for30 second period). Has anyone seen this problem? I have the server inAUDIT FAILUREs only. The server has become very unresponsive, things thattook 3 seconds now take over 15 seconds.Any ideas???
I need to track all changes to my database at the application level. Say if anybody logs into the application with his user id and password, any add,delete,update should be tracked. This is an application with 5000 transactions a day. I thought of using a trigger on all the tables to capture the following fields.
[UserID]
[AuditDate]
[TableName]
[PrimaryKey]
[Action]
,PreviousValue]
[CurrentValue]
I thought of adding all the application user as a database user. So then userID will be SYSTEM_USER , and AuditDate will be GETDATE(). Is this a good idea?
I dono whether the field primarykey is really essential in my audit table? Please advice.
The field 'Action' will be update/delete/insert. Previous and Current value will be the entire record from the temp tables inserted and deleted in triggers. I dono whether this is a good idea to store the entire record or just the field which was changed or updated(as updates r gonna be more than insert and delete), and the entire record for insert and delete. Is that possible? We have columns_updated() in sql server 2005 where we can we find the columns which was updated. Will that work?
Can i capture all actions(inser/update/delete) in a trigger with FOR INSERT,UPDATE,DELETE ? If so, how can i update "ACTION" and implement something different for update(as discussed earlier) . Is there any other field which you think i should capture which might be important as a part of auditing or tracking.
Can we use trigger for all tables in the database at one time?
I know there are too many questions . I wud greatly appreciate any thoughts or help for any portion of it.
I need to be able to audit the command Insert - Update - Delete sent to a table. I have attempted with a trigger but I have not been able to obtain the fields that have been upgraded.
It is possible to generate in the trigger the command Update - Insert - Delete sent?
Can anyone help, I am able to create a trigger that will populate a audit table everytime one of my tables columns data changes, but I have an applications from another user that has a stored proceudre and when that is called from an application it hit the original table twice, so the audit table will get a duplicate entry. How do you prevent an AUDIT TABLE from inserting a duplicate entry Here is my trigger:Create TRIGGER tg_audit_task_order_awardees on xxxfor INSERT,UPDATE,DELETE as INSERT INTO audit_task_order_awardees( audit_log_type, to_awardee, solicitation_id, contract_id, order_number, amount, show_public, audit_changedatetime, audit_user) Select 'OLD', del.to_awardee, del.solicitation_id, del.contract_id, del.order_number, del.amount, del.show_public, getdate(), del.modified_user FROM deleted del
/* for a new record */ INSERT INTO audit_task_order_awardees( audit_log_type, to_awardee, solicitation_id, contract_id, order_number, amount, show_public, audit_changedatetime, audit_user) Select 'NEW', ins.to_awardee, ins.solicitation_id, ins.contract_id, ins.order_number, ins.amount, ins.show_public, getdate(), ins.modified_user FROM inserted ins
I'm trying to create and Audit report for different tables.I'd like to use an generic table structure for the report which will be temp.I was thinking something like this:
I've already created the audit table with same structure plus all field duplicated and action type. (this was a request)Now I need to group it for the report.following an example of audit table I've created
tblTest_audit ( ID ID_new Name Name_new Address Address_new )
We would like to create a audit table for storing information if anyobjects like tables / stored procedures are dropped.Table would also contain information about time and the user name whohas dropped the object.Any help would be great.Thanks in advance.Kamal
