« first day (1481 days earlier)      last day (3387 days later) » 

12:54 AM
4
Q: Open source tools for Oracle schema compares

MacGyverI'm looking for an open source tool to generate DDL scripts. Many tools call these synchronization scripts. We essentially want to compare schema of different environments (ie: DEV to QA, QA to PROD) to manage object migrations/deployments a little easier. Do open source tools exist like this ...

The accepted answer for that annoys me because of the question clearly stating open source. The answer costs lots of money.
 
 
14 hours later…
3:22 PM
@ConcernedOfTunbridgeWells Nice one. I think existentialist predicates would be ones that cause a query optimizer to fall into a state of ennui and self-doubt and ponder the universe of all possible execution paths.
 
 
4 hours later…
7:49 PM
@JackDouglas I think the query should be SELECT * FROM table WHERE col2 >= val1 AND col2 <= val2 **AND col1 >= val1** ;
 
@ypercube I wondered about that but val1<>val1
don't worry I know what you mean
 
But I'm not sure. I have to reread th econdition
 
I think the extra condition is implied
 
Doesn't his condition mean that if we select all the rows and order by col1 then col2 values will be ordered the same way?
If yes, then your query is not equivalent.
Neither is the query I posted above
Oh, I see what you are doing now.
WHy don't you write it in one query - instead of 3?
 
@ypercube yes, you can do that
I should have added that at the end
I had it split out to demo the steps
(and to get simpler plans)
 
8:05 PM
select bar
from t
where foo >= 500
order by foo, bar
limit 1
(the first 2 steps in one)
 
that's good
marginally more efficient too
The OP might not have thought of that - you should post it
 
You can include it in your answer. It's not that different from yours.
And if the columns are unique, you only need order by foo in the first
 
you only need order by foo even if they aren't unique I think?
wait, no
that's rubbish
 
No, only if they are unique.
 
I'm probably not going to change mine, but I'd vote for a new answer from you :)
 
8:12 PM
well, at least one of them being unique is ok
@JackDouglas No time for answers, I'm, watching Star Trek season 6.
 
especially if you come up with a nice way of showing val1/2 without hard-coding
@ypercube :)
star, favourites, later :)
you might be interested to know that ypercube suggested a simpler version of this answer in The Heap — Jack Douglas ♦ 8 secs ago
 

« first day (1481 days earlier)      last day (3387 days later) »