Microsoft SQL Server forum - Multiple Insertions

Author

Multiple Insertions

Mark

2005-06-27, 11:23 am

Hi everyone,

I just need a bit of advice as to where to start tackling a problem, if
thats possible - thanks very much.

I need a single stored procedure to make several inserts into my msde
database. There will be two arguments to the stored proc. The first
is a title argument which needs to be inserted into the first table
after which the autonumbered primary key is captured with @@identity.

The second argument is a delimited list of foreign keys which need to
be inserted into the second table along with the new key from the first
statement. This table is a link table with two columns - both foreign
keys - ie its the link table in a many to many relationship.

My problems is that as far as I know I can't use arrays in sql server
cause it doesn't support them. And this has come about because I don't
know how many rows need to be inserted into the link table. But there
will always be at least one.

I know I need to do this in a loop, but how do I split up the the
second argument so that I can?

Thanks,

Mark

Hugo Kornelis

2005-06-27, 1:23 pm

On 27 Jun 2005 08:22:10 -0700, Mark wrote:

>
>Hi everyone,
>
>I just need a bit of advice as to where to start tackling a problem, if
>thats possible - thanks very much.
>
>I need a single stored procedure to make several inserts into my msde
>database. There will be two arguments to the stored proc. The first
>is a title argument which needs to be inserted into the first table
>after which the autonumbered primary key is captured with @@identity.
>
>The second argument is a delimited list of foreign keys which need to
>be inserted into the second table along with the new key from the first
>statement. This table is a link table with two columns - both foreign
>keys - ie its the link table in a many to many relationship.
>
>My problems is that as far as I know I can't use arrays in sql server
>cause it doesn't support them. And this has come about because I don't
>know how many rows need to be inserted into the link table. But there
>will always be at least one.
>
>I know I need to do this in a loop, but how do I split up the the
>second argument so that I can?
>
>Thanks,
>
>Mark

Hi Mark,

Check out this site for a wealth of possible solutions:

http://www.sommarskog.se/arrays-in-sql.html

Best, Hugo
--

(Remove _NO_ and _SPAM_ to get my e-mail address)

--CELKO--

2005-06-27, 8:23 pm

>> I just need a bit of advice as to where to start tackling a problem, if that is possible - thanks very much. <<

You need to go back to foundations. You missed the major points of
RDBMS and are trying to write a 1950's file system in SQL.

1) Autonumbering is totally non-relational and cannot be a key by
definition. This is foundations, not fancy stuff. A key is a subset
of attributes that makes a row unique within a table; it has to do with
the data model and not the current state of the hardware on which the
data is stored.

2) An INSERT INTO statement works on one and only one base table.

3) There are only scalar value parameters; there are no lists, arrays,
etc. There are a bunch of kludges where you write a parser in T-SQL,
if you do not care about maintaining or porting your code.

4) There is no such term as "link table" -- link is a term from
navigational databases and assembly language. It is a many-to-many
relationship.

5) We do not like to write procedural code, so you should avoid loops.

6) You do not insert keys into a table; you insert rows. WHOLE rows.
You probably have more (non-key) columns in the second table to fill
in.

7) Please post DDL, so that people do not have to guess what the keys,
constraints, Declarative Referential Integrity, data types, etc. in
your schema are. Sample data is also a good idea, along with clear
specifications. It is very hard to debug code when you do not let us
see it.

Your code should look something like this, assuming a vanilla invoice
schema:

BEGIN
INSERT INTO Invoices (invoice_nbr, ..)
VALUES (@new_invoice_nbr, .. );

INSERT INTO InvoiceDetails (invoice_nbr, sku..)
SELECT @new_invoice_nbr, sku..
FROM WorkingTable;
END;

Invoice_nbr should have a CHECK() constraint to validate it, of course.

Robin Tucker

2005-06-28, 7:23 am

This is the version of split array return table from sommarskog. I use this
to pass in arrays of keys (space separated). It's more efficient than
populating a temporary table with keys programmatically every time you need
to join (as Celko seems to suggest). Although in general Celko is correct
from a purest point of view, I do believe that a small function like this
allowing you to pass and split arrays in a stored procedure has more utility
than it does downsides. For example, my client needs to periodically check
a set of rows to see if they have been changed. These records are in no
particular order (whichever rows the user happens to be viewing). Instead
of writing each one to a working table one by one and then executing an SP
to check their timestamps, I pass in an array of keys, split it and join
with the split table to return the update state.

However, in your example, perhaps a working table would be a better idea. I
would only advise using array splitting algorithms server-side if they are
just autonumber unique keys, rather than whole rows of information. It's
just a quickish method for fetching arbitrary rows from your tables.

CREATE FUNCTION dbo. func_Split_Array_Ret
urn_Table (@list NTEXT)
RETURNS @Table TABLE ( listpos INT IDENTITY(1, 1) NOT NULL, number INT NOT
NULL) AS

BEGIN
DECLARE @pos INT, @textpos INT, @chunklen SMALLINT, @str NVARCHAR(4000),
@tmpstr NVARCHAR(4000), @leftover NVARCHAR(4000)

SET @textpos = 1
SET @leftover = ''

WHILE @textpos <= datalength(@list) / 2
BEGIN

SET @chunklen = 4000 - datalength(@leftover
) / 2
SET @tmpstr = ltrim(@leftover + substring(@list, @textpos,
@chunklen))
SET @textpos = @textpos + @chunklen
SET @pos = charindex(' ', @tmpstr)

WHILE @pos > 0
BEGIN

SET @str = substring(@tmpstr, 1, @pos - 1)

INSERT @Table (number)
VALUES
(convert(int, @str))

SET @tmpstr = ltrim(substring(@tmp
str, @pos + 1, len(@tmpstr)))
SET @pos = charindex(' ', @tmpstr)

END

SET @leftover = @tmpstr
END

IF ltrim(rtrim(@leftove
r)) <> ''
INSERT @Table (number)
VALUES
(convert(int, @leftover))

RETURN

END

"Mark" <mjmather@gmail.com> wrote in message
news:1119885730.339475.84280@g47g2000cwa.googlegroups.com...
>
> Hi everyone,
>
> I just need a bit of advice as to where to start tackling a problem, if
> thats possible - thanks very much.
>
> I need a single stored procedure to make several inserts into my msde
> database. There will be two arguments to the stored proc. The first
> is a title argument which needs to be inserted into the first table
> after which the autonumbered primary key is captured with @@identity.
>
> The second argument is a delimited list of foreign keys which need to
> be inserted into the second table along with the new key from the first
> statement. This table is a link table with two columns - both foreign
> keys - ie its the link table in a many to many relationship.
>
> My problems is that as far as I know I can't use arrays in sql server
> cause it doesn't support them. And this has come about because I don't
> know how many rows need to be inserted into the link table. But there
> will always be at least one.
>
> I know I need to do this in a loop, but how do I split up the the
> second argument so that I can?
>
> Thanks,
>
> Mark
>

Sponsored Links