T-SQL - Is there a (free) way to compare data in t

Question

I have table a and table b. (SQL Server 2008)

Both tables have the exact same schema.

For the purposes of this question, consider table a = my local dev table, table b = the live table.

I need to create a SQL script (containing UPDATE/DELETE/INSERT statements) that will update table b to be the same as table a. This script will then be deployed to the live database.

Any free tools out there that can do this, or better yet a way I can do it myself?

I'm thinking I probably need to do some type of a join on all the fields in the tables, then generate dynamic SQL based on that.

Anyone have any ideas?

EDIT

Okay, thought I'd clarify this question a little.

The table I need to synchronize is a simple look-up table. The data is very simple and straightforward.

Here's an idea of what TABLE A might look like:

IdFoo          Activity      IsFoo
1              Foo           1
2              Bar           0

Here's an idea of what TABLE B might look like:

IdFoo          Activity      IsFoo
1              Foo           1
2              Bar           1

Basically, all I want to do is update that BIT column (IsFoo) in TABLE B to match the corresponding value in TABLE A for the same IdFoo.

Keep in mind:

• TABLE A is on my local machine
• TABLE B is on the live server

Obviously I have a (reliable) backup of TABLE B on my local machine which i need to script against, then run the script on the live server.

The table also has referential integrity (other columns I didn't show). Which means I can't just delete everything in TABLE B and do an insert from TABLE A.

This script will be a once off. So no need to do things like linked server, replication, etc. Appreciate the answers though guys. =)

EDIT:

Ok - so I've gone with Oleg's answer (VS2010 Data Compare). Quick, easy, and works a charm.

Not to say the other answers are incorrect. I appreciate all the answers!

Answer 1

Since it is one off, you can use this query to find rows that are different in this two tables:

(SELECT * FROM TABLE_A
 MINUS
 SELECT * FROM TABLE_B)

 UNION ALL

(SELECT * FROM TABLE_B
 MINUS
 SELECT * FROM TABLE_A)

MINUS will compare records field by field, then it will discard records from first table for which there is identical record in second table. This works like this:

• First MINUS gets all records from TABLE_A that are not in TABLE_B
• Second MINUS gets all records from TABLE_B that are not in TABLE_A
• Union gets all records from both tables for which there is no matching record in other table.

Now you can insert those records in some temp table and then do inserts/updates.

Depending on your needs, you can restrict field list for comparing.

Note that you need primary key for this to work.

Edit:
Ooops. SQL Server does not support MINUS operator. I'm working with ORACLE last year and half, so this was automatic.

You can use EXCEPT operator instead. See this article: EXCEPT and INTERSECT (Transact-SQL)

Edit 2:

Re scherand's comment:
If he really cannot connect from local machine to live server, than he can simply dump TABLE_A and load it on server. One way or the other, goal is to change table data on live server.

Answer 2

You can also try using the import and export data provided by SQL Server 2008. Its is fairly straight way to copy all the data from anywhere to anywhere. I do the same thing and works perfectly.

Answer 3

what happens if you:

1. copy devTableA onto prod,
2. suspend dependencies on prodTableB
3. rename prodTable B to prodTableB1,
4. rename devTableA to prodTableB
5. reimplement dependencies on new prodTableB
6. test new prodTableB and then delete out-named old prodTableB

?

or if you have to be more respectful of the existing prodTableB and your identity column values in both dev and prod match (even if the values are not continuous)...

declare @iLast int, @x int, @y int
select @iLast = (select MAX(id) from prodTableB)
set @x = 1
while @x <= @iLast
begin
  select @y = (select COUNT(*) from prodTableB where id = @x)
  if @y = 1
    begin
      update prodTableB set isFoo = (select isFoo from devTableA where id=@x
    end
  @x=@x+1
end

Answer 4

In very simple cases you will be able to define in your local SQL Server a new Linked Server (see in Microsoft SQL Server Management Studio under "Server Objects" / "Linked Server") and use INNER JOIN and OUTER JOIN to find out differences between A and B tables.

In a real and more complex situation you should take into consideration Referential Integrity, different Foreign Keys, and Identity (auto incremental) fields existing in the destination database, so the update script will be more complex. So I'll recommend you don't spend your time in creating synchronization of your developer and production DB and use a standard tool instead. I use for example the features of Visual Studio Team Edition 2008 (or Database edition) or Visual Studio 2010 Ultimate edition to compare data in two databases. It works very well.

Answer 5

You could use a data script generator that creates script for the inserts, and then use a file comparison tools such as WinMerge to compare the files to find the differences. There's an article on generating the data scripts on code project: http://www.codeproject.com/KB/database/sqlinsertupdategenerator.aspx

Answer 6

Late answer but can be useful to the thread visitors

Besides other mentioned solutions, I can suggest trying ApexSQL Data Diff. It can compare and sync SQL Server database data (in live databases as well as backups), and automate and schedule data migrations. It works with huge databases also, and is able to perform comparisons directly from SSMS.

You can download this tool for free and play with it. It features a fully functional free trial, and offers a community edition (after the trial period is over) that works on SQL Express and Azure SQL Database.

To learn more about the tool, you may visit http://www.apexsql.com/sql_tools_datadiff.aspx