After reading all this about Red-Gate I thought I’d give it a go.
It doesn’t work for me and I’m wondering how it can possibly work for anyone else with a database that has any interrelated objects (dependencies) at all.
I noticed the review was made on a database with only 3 tables, 2 views, 1 stored procedure, 2 database roles, 2 users, 1 default and only 2 user defined data types that I can see.
That to me is a meaningless test.
I’m a release manager at NTL in Hook, Hamshire, United Kingdom.
I ran a test on what used to be our most problematic database.
This database consists of the following objects: 796 tables (290 of which contain static data consisting of 107,186 rows), 62 views, 3352 stored procedures, 4 rules, 4 defaults, and 160 user defined functions.
That to me is a meaningful test.
I compared this to the database to be released and came up with 2609 schema changes and 169 tables with data changes consisting of 1603 deletes, 39379 inserts and 11409 updates making row changes 52391.
SQL Compare failed with 302 errors. After that I didn't bother with the data comparison.
Granted it is still better than working this out by hand, but we use another system which for our last 4 releases has never failed and does not need any modifications.
When people say this application fast (SQL Compare) – you have to put it into perspective.
I have to correct all these errors to release the software.
That would take days.
As I have said earlier, this was our most problematic database.
We now use DB Ghost which handles the release perfectly and integrates with our daily build process.
These guys are incredibly helpful and although are nowhere near as big as some of the other software vendors that say they compare and synchronise databases, they’re the best.
I can’t see how the other vendors say they ‘synchronise’ when clearly to me they don’t.