Showing posts with label dynamic. Show all posts
Showing posts with label dynamic. Show all posts

Sunday, February 12, 2012

Best way to do a dynamic bulk insert to a table

My current project is creating a social network for the university I work for. One of the features allows members of a group to send a message to all other group members. Currently, I run a foreach loop over each of the group members, and run a separate INSERT statement to insert a message into my messages table. Once the group has several hundreds members, everybody starts getting timeout errors. What is the best way to do this?

Here are two suggestions I've received: construct one sql statement that would contain multiple INSERT statements. It would be a large statement like:

INSERT into [messages] (from_user, to_user, subject, body) VALUES (@.from_user, @.to_user, @.subject, @.body); INSERT into [messages] (from_user, to_user, subject, body) VALUES (@.from_user2, @.to_user2, @.subject2, @.body2); INSERT into [messages] (from_user, to_user, subject, body) VALUES (@.from_user3, @.to_user3, @.subject3, @.body3);

etc...

Or, do the foreach loop in a stored procedure. I know the pros and cons of sprocs versus dynamic sql is a sticky subject, and, personally, I'd prefer to keep my logic in the C# code-behind file. What is the best way to do this is an efficient manner? I'd be happy to share some code, if that would help. Thanks for your input!

I think what you want is something like this. 1 single query.

INSERT into [messages] (from_user, to_user, subject, body)
SELECT @.from_user,userid,@.subject,@.body FROM groupmembers WHERE groupid=@.groupid AND userid<>@.from_user

Your input parameters:

@.from_user = the sending user
@.subject = the subject
@.body = the body
@.groupid=the recieving groupid


|||

If you going to insert data to a table that alreay has some data, I suggest you to use whatgunteman wrote.

If not, I suggest you to use:

1SELECT *2INTO MyNewTable-- it will be created automatically here3FROM MyTable1 t14INNERJOIN5 MyTable2 t26ON (t1.rid = t2.rid)7

Good luck.

|||

Thanks that worked like a charm. I can't believe I never thought of putting a subquery in an insert statement.

Best way to create dynamic update statement

In general, What is the best approach in creating a dynamic update
stored procedure, that can handle recieving varying input paramters and
update the approporiate columns.Depends on the requirements but one possibility is to use NULL
parameters to represent values that shouldn't be changed:

UPDATE YourTable
SET col1 = COALESCE(@.col1, col1),
col2 = COALESCE(@.col2, col2),
col3 = COALESCE(@.col3, col3)
... etc
WHERE ...

--
David Portas
SQL Server MVP
--|||>> In general, What is the best approach in creating a dynamic update
stored procedure, <<

In general, building dynamic is a bad idea. It says that you don't
know what you are doing, so you are turning over control of the system
at runtime to any random user, present or future. SQL is a compiled
language, not like BASIC.|||what would be wrong with using:
UPDATE YourTable
SET col1 = COALESCE(@.col1, col1),
col2 = COALESCE(@.col2, col2),
col3 = COALESCE(@.col3, col3)
... etc
WHERE ...

if i want to have one stored procedure to update a table.|||jw56...@.gmail.com wrote:
> if i want to have one stored procedure to update a(ny) table.

what would be wrong

--Strider|||jw56...@.gmail.com wrote:
> if i want to have one stored procedure to update a(ny) table.

what would be wrong

--Strider|||I think this is just some confusion over terminology. The term "dynamic
update" or "dynamic code" refers to code that references metadata
(usually table and column names) dynamically - elements of the code
being constructed at runtime. This is not generally good practice for
various reasons to do with performance, security, maintainability and
modular design. In your case however, no dynamic code is necessary.

--
David Portas
SQL Server MVP
--