Combine tables from 2 SQL servers with different Schemas and update as new data is entered

Question

 I have 2 SQL server 2000 machines, I need to take a table from each one and combine them together based on a date time stamp.  The first machine has a database that records information based on an event it is given a timestamp the value of variable is stored and a few other fields are stored in Table A.  The second machine Table B has test data entered in a lab scenario.  This is a manufacturing facility so the Table A data is recorded by means of a third party software.  Whenever a sample is taken in the plant the event for Table A is triggered and recorded in the table.  The test data may be entered on that sample in Table B several hours later the lab technician records the time that the sample was taken in Table B but it is not exact to match with the timestamp in Table A.  I need to combine each of these tables into a new SQL server 2005 database on a new machine.  After combining the tables which I am assuming I can based on a query that looks at the timestamp on both Tables A & B and match the rows up based on the closest timestamp. I need to continuously update these tables with the new data as it comes in.  I havent worked with SQL for a couple of years and have looked at several ways to complete this task but havent had much luck.  I have researched linked servers, SSIS, etc Any help would be greatly appreciated.

Answer 1

Hi Catalyst,
Welcome to the MSDN Forums!
I would have a concern with respect to your issue.  If there is no way to identify a sample, save for the closest timestamp on table A - what would happen if two samples were taken in the plant, and the test results were entered in reverse order?  The samples would not tie back to the appropriate measurements.  Can't you get an identifier of some sort to tie that sample back to table A?
If you can't, something like this should work, but I'm not sure how you'll "serialize" the samples to avoid the situation above.
Code Snippet
--Create sample data
DECLARE @TableA TABLE ( SampleNo int NOT NULL,
SampleTime datetime);
DECLARE @TableB TABLE ( SampleTakenTime datetime,
SampleColour varchar(50),
SampleQuality char(1),
SampleResultNo int);
INSERT @TableA SELECT 1, GETDATE()
WAITFOR DELAY '00:00:10';
INSERT @TableB SELECT GETDATE(), 'Orange', 'A', '5'
WAITFOR DELAY '00:00:20';
INSERT @TableB SELECT GETDATE(), 'Purple', 'D', '7'
WAITFOR DELAY '00:00:10';
INSERT @TableA SELECT 2, GETDATE()
--Actual Query begins now
SELECT *
FROM @TableA AS a
JOIN @TableB AS b ON ABS(DATEDIFF(millisecond,a.SampleTime,b.SampleTakenTime)) = (SELECT MIN(ABS(DATEDIFF(millisecond,ta.SampleTime,tb.SampleTakenTime)) )
FROM @TableB tb
JOIN @TableA ta ON a.SampleNo = ta.SampleNo)
Now, to account for the fact that you are querying data from a different server, you need to create a linked server (see sp_addlinkedserver and sp_addlinkedsrvlogin) and then use the fully qualified name (server.database.schema.table) to reference the table.
Hope this helps,
Aaron