Database

Re: ado.net connections and .dispose?

30-Aug-04 1:27

Which means that your MS Press book is wrong. I never said I was right, I just said "If I Recalled Correctly"

"If a man empties his purse into his head, no man can take it away from him, for an investment in knowledge pays the best interest." -- Joseph E. O'Donnell

Not getting the response you want from a question asked in an online forum: How to Ask Questions the Smart Way!

Rocky Moore31-Aug-04 20:01

Rocky Moore

31-Aug-04 20:01

Sorry, I did not read this earlier.. Just wrote up a tip about it yesterday as someone else asked me the question about a week ago. I did not remember correctly and did some research. Come to the same conclusion as the thread here. Many people have argued that the dispose actually would not allow it to be used in the connection pool. Who is right and wrong? In the use net groups there were people who worked on the code that said absolutely that there was little difference. About the only thing the dispose did that close did not was delete the connection string.

Here is more info along with preventing a leak:

http://www.hintsandtips.com/ShowTip/129/hat.aspx[^]

Rocky <><
www.HintsAndTips.com - Now with GMail Queue
www.MyQuickPoll.com - 2004 Election poll is #33
www.GotTheAnswerToSpam.com

"We plan for the future, we learn from the past, we live life in the present!"

Identity primary key

Diego F.26-Aug-04 23:18

Diego F.

26-Aug-04 23:18

Hi. I have a table with a primary key that I'd like to be an identity, so I don't worry about giving it a value manually (just increases automatically. 1, 2, 3, ...)

The problem is that I have another table where these key is a foreign key, so in my application, I must fill the value manually to the second table, but then I need to get the value in the first table.

I mean, imagine I have a table with: (id, name, age), where 'id' is the primary key and is identity, with values 1,2,...

Then I have another table with: (id, phone), where 'id' is the foreign key.

In my application, I can fill entries in the first table without giving values to id, as it is done automatically, but in the second table, I must give manually the values. How can I do that?

Should I make a query to get the 'id'?

Regards,

Diego F.

David Salter26-Aug-04 23:28

David Salter

26-Aug-04 23:28

You can get an the last inserted identity (on SQL Server) by calling

select @@identity

You can then insert this into your second table.

store your internet favourites online - www.my-faves.co.uk

Diego F.26-Aug-04 23:50

Diego F.

26-Aug-04 23:50

Thanks. I was searching about this and I found also SCOPE_IDENTITY(). Apart from using one or the other, how can I get the value in my application.

I mean, I saw people using: "Select @value = scope_identity()", but where do I catch the @value variable?

Regards,

Diego F.

David Salter27-Aug-04 1:12

David Salter

27-Aug-04 1:12

If you do a select @@identity, you get the results into your application exactly the sameway that you do with any other select statement.

With select @@identity, you do not need to select the value into a variable.

store your internet favourites online - www.my-faves.co.uk

Colin Angus Mackay27-Aug-04 1:33

SQL and the growing log file

27-Aug-04 1:33

DECLARE @id int;
INSERT INTO aTable(name, age)
VALUES('Joe Bloggs', 21);
SELECT @id = @@IDENTITY;
INSERT INTO anotherTable(id, phone)
VALUES(@id, '212-555-2468');

If you are doing this from within an ASP.NET application then you can do something like

CREATE PROCEDURE InsertATable
(
    @name varchar(32),
    @age int
)
AS
INSERT INTO aTable(name, age)
VALUES('Joe Bloggs', 21);
SELECT @@IDENTITY as Id;
GO

SqlCommand cmd = new SqlCommand("InsertATable", connection);
cmd.Parameters.Add("@name", "Joe Bloggs");
cmd.Parameters.Add("@age", 21);
cmd.CommandType = CommandType.StoredProcedure;
int id = (int)cmd.ExecuteScalar();
// I now have the id of the newly created record in my .NET application.

Does this help?

totig25-Aug-04 20:02

totig

25-Aug-04 20:02

I have a log file that has grown a lot. I need to get it smaller. I have been reading SQL books online, and while it explains the problem, I cannot work out what the solution is to it. The database is being backed up each night, not the log. I would like to know if there is a way that the log can drop all the logs until the last database backup.

mikasa26-Aug-04 3:40

mikasa

26-Aug-04 3:40

If you're not concerned with the Log, change it to "Simple" mode on the Server. I am pretty sure though that if you are backing up the Logs, it will shrink the Log, but I can't be sure off the top of my head.

Mekong River26-Aug-04 3:41

Mekong River

26-Aug-04 3:41

You should check the property of SQL server database. In the property dialog box click on the third tab "Transaction Log" property. This page contain the information about setting how the log file growth it size. If you have a problem with the size of the log file growth very fast try to choose an option "Restrict File Growth" option and set it value.

Hope it is work!!!

APO-CEDC
Save Children Norway-Cambodia Office

totig26-Aug-04 3:47

totig

26-Aug-04 3:47

I found that changing the db to simple, and then running the shrink db command, the log file has dropped and wont grow again. Limiting the log in this instance would not have worked, as the db was expecting the log file to be backed up as well. Had the log file been set to backup, the log files size would have decreased.

Nested Transaction between ADO.NET and T-SQL

Rocky Moore31-Aug-04 20:18

Rocky Moore

31-Aug-04 20:18

Glad you brought this up as I reminded me to make a tip out of the backup procedure I use on my server.

http://www.hintsandtips.com/ShowTip/130/hat.aspx[^]

I use the "backup log with truncate_only" and then use the "DBCC SHRINKFILE".

If you restrict the length of the log, it will fill your NT Event log with error messages when the log reaches the limit.

Rocky <><
www.HintsAndTips.com - Now with GMail Queue
www.MyQuickPoll.com - 2004 Election poll is #33
www.GotTheAnswerToSpam.com

"We plan for the future, we learn from the past, we live life in the present!"

Sanjeev Kumar25-Aug-04 14:29

25-Aug-04 14:29

I am writing an application that calls numbers of stored procedure wrapped in an ADO.NET transaction. Some of the stored procedure implements T-SQL transactions.
Is it safe to call Stored Procedure like this or there is potential of data corruption.

Have anyone seen any document or article that explain how this works.

Sample code

using (SqlConnection conn = new SqlConnection(connectionString))
{
conn.Open();
SqlTransaction theTran = conn.BeginTransaction();
try
{
// create command object
// set Command Parameters
// Set Command Parameters values
// EXECUTE Stored Procedure 1

// create command object
// set Command Parameters
// Set Command Parameters values
// EXECUTE Stored Procedure 2

....

theTran.Commit();
}
catch
{
theTran.Rollback();
}
conn.Close();
}

Stored Procedure

Create Procedure InsABC
@NAME varchar(50),
@Address varchar(100)
as
BEGIN
BEGIN TRANSACTION
....
-- Do INSERT ...
....

-- In case of error rollback

...
COMMIT TRANSACTION

END

S Sansanwal25-Aug-04 16:33

S Sansanwal

25-Aug-04 16:33

Why don't you create a single StoreProcedure that is called by client.
Inside this SP, call all other procedures.

Sanjay Sansanwal
www.sansanwal.com

Sanjeev Kumar25-Aug-04 16:46

25-Aug-04 16:46

No, I can't combine all the SP, some times same SP is called multiple times depending on Data. Its all dynamic based on data which comes from client application

Colin Angus Mackay25-Aug-04 22:34

25-Aug-04 22:34

This isn't a problem. You can nest transactions inside each other. Just make sure that they all get committed or rolledback properly otherwise you might get errors saying that the transaction count before starting some SP is 1 and afterwards is 2 (because you forgot to commit or rollback the transaction before the SP exited)

Sanjeev Kumar26-Aug-04 7:07

26-Aug-04 7:07

Thanks Colin,
I have written an application just as proof of concept and it works fine. But, my boss think its not safe to do like this and he refering to Online Book Which say :-
----------------------
Specifying Transaction Boundaries
You can identify when SQL Server transactions start and end with Transact-SQL statements or API functions and methods.

Transact-SQL statements

Use the BEGIN TRANSACTION, COMMIT TRANSACTION, COMMIT WORK, ROLLBACK TRANSACTION, ROLLBACK WORK, and SET IMPLICIT_TRANSACTIONS statements to delineate transactions. These are primarily used in DB-Library applications and in Transact-SQL scripts, such as the scripts that are run using the osql command prompt utility.

API functions and methods

Database APIs such as ODBC, OLE DB, and ADO contain functions or methods used to delineate transactions. These are the primary mechanisms used to control transactions in a SQL Server application.

Each transaction must be managed by only one of these methods. Using both methods on the same transaction can lead to undefined results. For example, you should not start a transaction using the ODBC API functions, and then use the Transact-SQL COMMIT statement to complete the transaction. This would not notify the SQL Server ODBC driver that the transaction was committed. In this case, use the ODBC SQLEndTran function to end the transaction.--------------

As per this document I am not suppose to mix transaction between ADO.NET and T-SQL. And I am not mixing transactions instead, I am nesting it.
Is there is any published article or document online which can help me explaning to my BOSS?

Thanks
Sanjeev

Colin Angus Mackay26-Aug-04 8:45

26-Aug-04 8:45

What the book is saying, and you've already realised, is that if you start a transaction in ADO.NET you must complete it in ADO.NET. If you start a transaction in T-SQL you must complete it in T-SQL. You can of course nest transactions inside each other.

For example (in this pseudo code)

Start transaction 1
  Start transaction 2
    -- Do stuff
  End transaction 2
End transaction 1

The above is perfectly fine. Transaction 1 and 2 can both be in T-SQL, or both be in ADO.NET or Transaction 1 can be in ADO.NET with Transaction 2 being in T-SQL.

Think of it like loops, you can start a for loop and call into another method that perhaps has another for loop inside it. The for loop in the inner method must complete before the for loop in the outer method can complete. Does this make sense?

As for a published article... I am not sure about that. The text you've quoted seems perfectly reasonable to me. How competant is your boss? (Hmmm... maybe that is a bad question) How technical is your boss? Could he understand it in a way similar to the for loop analogy? If he is less technical, how about a Russian doll analogy? (The Russian dolls that are hollow and one fits inside the other, they split at the waist so you can enclose one doll around another. Pretend that the base is a start transaction, the head is the end transaction, and each size of doll is a different technology. You must obviously match the correct base with the correct head otherwise they won't fit together properly.)

Does this help?

Sanjeev Kumar26-Aug-04 13:59

26-Aug-04 13:59

Thanks for the reply
I Hope my boss will understand this.

Colin Angus Mackay27-Aug-04 6:07