How to determine internal name of table-valued variable in MS SQL Server 2005 - tsql

Question

The name of a temporary table such as #t1 can be determined using
select #TableName = [Name]
from tempdb.sys.tables
where [Object_ID] = object_id('tempDB.dbo.#t1')
How can I find the name of a table valued variable, i.e. one declared by
declare #t2 as table (a int)
the purpose is to be able to get meta-information about the table, using something like
select #Headers = dbo.Concatenate('[' + c.[Name] + ']')
from sys.all_columns c
inner join sys.tables t
on c.object_id = t.object_id
where t.name = #TableName
although for temp tables you have to look in tempdb.sys.tables instead of sys.tables. where do you look for table valued variables?
I realize now that I can't do what I wanted to do, which is write a generic function for formatting table valued variables into html tables. For starters, in sql server 2005 you can't pass table valued parameters:
http://www.sqlteam.com/article/sql-server-2008-table-valued-parameters
moreover, in sql server 2008, the parameters have to be strongly typed, so you will always know the number and type of columns.

Answer 1

Table variable metadata is viewable in tempdb.sys.tables too. This is easily verifiable from the below
declare #t2 as table ( [38F055D8-25D9-4AA6-9571-F436FE] int)
SELECT t.name, t.object_id
FROM tempdb.sys.tables t
JOIN tempdb.sys.columns c
ON t.object_id = c.object_id
WHERE c.name = '38F055D8-25D9-4AA6-9571-F436FE'
Example Results
name object_id
------------------------------ -----------
#4DB4832C 1303675692
But you will notice the object name is auto generated and bears no relation to the variable name.
If you do not have a guaranteed unique column name that you can use to filter on as above and the table variable has at least one row in it you can (from SQL Server 2008 onwards) use %%physloc%% and DBCC PAGE to determine this information. Example below.
DECLARE #t2 AS TABLE ( a INT)
INSERT INTO #t2
VALUES (1)
DECLARE #DynSQL NVARCHAR(100)
SELECT TOP (1) #DynSQL = 'DBCC PAGE(2,' + CAST(file_id AS VARCHAR) + ',' +
CAST( page_id AS VARCHAR) +
',1) WITH TABLERESULTS'
FROM #t2
CROSS APPLY sys.fn_PhysLocCracker( %% physloc %% )
DECLARE #DBCCPage TABLE (
[ParentObject] [VARCHAR](100) NULL,
[Object] [VARCHAR](100) NULL,
[Field] [VARCHAR](100) NULL,
[VALUE] [VARCHAR](100) NULL )
INSERT INTO #DBCCPage
EXEC (#DynSQL)
SELECT VALUE AS object_id,
OBJECT_NAME(VALUE, 2) AS object_name
FROM #DBCCPage
WHERE Field = 'Metadata: ObjectId'

Answer 2

From Books Online:
A table variable behaves like a local variable. It has a well-defined scope, which is the function, stored procedure, or batch in which it is declared.
Given this, there should be no need to look up this value at run-time because you have to know it at design-time.

Answer 3

I don't believe you can, as table variables are created in memory not in tempdb.

Answer 4

On the topic of passing arbitrary lists/arrays into a SQL Server 2005 function or sproc, the least hokey way I know is to use an XML variable. If desired, that XML variable can be a strongly typed XML type that is associated w/ an XML Schema.
Given a list passed into a procedure/function as XML, you can extract that list into a table variable or temp table via "shredding".
"To shred" XML means to transform in the opposite direction--from XML to rowset(s). (The FOR XML clause causes a rowset to XML transformation.)
In the user-defined table function
CREATE FUNCTION [dbo].[udtShredXmlInputBondIdList]
(
-- Add the parameters for the function here
#xmlInputBondIdList xml
)
RETURNS
#tblResults TABLE
(
-- Add the column definitions for the TABLE variable here
BondId int
)
AS
BEGIN
-- Should add a schema validation for #xmlInputIssuerIdList here
--Place validation here
-- Fill the table variable with the rows for your result set
INSERT #tblResults
SELECT
nref.value('.', 'int') as BondId
FROM
#xmlInputBondIdList.nodes('//BondID') as R(nref)
RETURN
END
if the #xmlInputBondIdList is an XML fragment of the expected structure like that immediately below and is invoked as follows
DECLARE #xmlInputBondIdList xml
SET #xmlInputBondIdList =
'<XmlInputBondIdList>
<BondID>8681</BondID>
<BondID>8680</BondID>
<BondID>8684</BondID>
</XmlInputBondIdList>
'
SELECT *
FROM [CorporateBond].[dbo].[udtShredXmlInputBondIdList]
(#xmlInputBondIdList)
the result will be the rowset
BondId
8681
8680
8684
A couple other examples can be found at http://forums.microsoft.com/MSDN/ShowPost.aspx?PostID=678284&SiteID=1