Viewing 15 posts - 91 through 105 (of 229 total)
I was too shy to express it so forcefully but yes I agree.
March 4, 2016 at 9:04 am
Sean Lange (3/3/2016)
March 4, 2016 at 8:19 am
"But for the sanity of folks who use database software, DON'T HAVE YOUR RDBMS VENDOR HARD WIRE MEANINGS INTO NULLS."
Speaking from bitter experience perhaps ?
February 29, 2016 at 1:29 pm
rstone (2/29/2016)
February 29, 2016 at 12:11 pm
Indeed, I forgot that I should have also added the metadata on top of each side table for each nullable field. An extra layer of complication yet.
February 29, 2016 at 8:47 am
Yes, my fix did imply one specific meaning for a null - if you cannot name the judge for instance it is because it has not been done yet. Cannot...
February 29, 2016 at 8:38 am
That doesn't get rid of NULLs. You'd have to LEFT JOIN to the other tables ... which produces NULL if you don't find a match!
Well you can always "fix"...
February 26, 2016 at 3:01 pm
kenambrose (2/26/2016)
You just take any column where you would otherwise need nulls, and make a related table for it and ...
February 26, 2016 at 2:46 pm
roger.plowman (2/26/2016)
j-1064772 (2/26/2016)
February 26, 2016 at 2:19 pm
An absolute ban on NULL can impose excessive constraints in real life. Events cannot all be reserved for after-the-fact data entry.
For instance, when drawing up a tentative schedule for court...
February 26, 2016 at 1:04 pm
JediSQL (10/31/2014)
The one big exception is a CHECK CONSTRAINT. With a CHECK CONSTRAINT the assertion just has to be not provably FALSE. Say we have [font="Courier New"]......
February 26, 2016 at 11:44 am
The execution plans do not tell the whole story. I have seen some cases by Gila Monster where the execution was simplified but the load on the server was not...
February 18, 2016 at 7:34 am
davidawest00 (2/17/2016)
You can do the same thing with a correlated sub-query, and they usually scale better than any CTE because a CTE...
February 17, 2016 at 11:57 am
Re: Alternative method for those without LEAD() and LAG()
The original code WITH LEAD() and LAG() seems easier to read for me.
The alternative method would apply in a version of SSQL...
February 17, 2016 at 9:42 am
redundant comment - deleted.
January 28, 2016 at 2:21 pm
Viewing 15 posts - 91 through 105 (of 229 total)