@AaronBertrand I took some of your advice regarding our transaction log testing. I split my table into 2 versions (current/archive), created a view and synonym to not break functionality, then implemented older merge syntax with - delete, update and insert. Not only did our transaction log backup on one table drop about 400MB but the performance improved. The procedure was taking about 480k+ms and I dropped it to about 15-20k+ms
an end user shouldn't be responsible for looking for security gaps
It's the jankyass suckholes that built the site using popups as navigation, defaulting to http for logins, etc, that should be hung from their entrails
PASS' problem is they do everything on the cheap - volunteers to build a web site that needs to stand up to the critique of people who build data solutions for a living.
And realistically, this is a web site for free training events. I think the general expectations are a lot higher than they need to be. Does it need to perfect? No.
@billinkc That's not correct. They don't just let anyone wave their hands and get source code.
I've sent far too many requests to HQ about things that don't work, that are broken, that perform poorly, impede our ability to do things on the sql saturday site
That's on our list. We've got a rewrite coming. It's going to address this
@mmarie I think they just had a lot of people who have no idea how to test for edge cases or for what real users would do. Did a search, it worked! Next? (Oh, I don't need to click back to make sure x is still there, etc)
Again, you get what you pay for. I could volunteer to go milk cows on a farm - I'd be helping, I suppose, but I probably won't get the same yield as someone with experience.
I just don't think they get the right volunteers, and I don't think they hire the right IT people, either. At one point I had 6 separate profiles associated with a single e-mail address. A MySQL newbie wouldn't let that happen, but PASS did. And it took them two weeks to figure out how to fix it.
North Bay is a city in Northeastern Ontario, Canada. It is the seat of Nipissing District, and takes its name from its position on the shore of Lake Nipissing.
== History ==
The site of North Bay was on the main canoe route west from Montreal (see Canadian Canoe Routes (early)). Apart from First Nations tribes, voyageurs and surveyors, there was little activity in the Lake Nipissing area until the arrival of the Canadian Pacific Railway (CPR) in 1882. The CPR started its westward expansion from Callander Station (later renamed Bonfield), Ontario; Bonfield was inducted into Canadian Railway Hall...
@AaronBertrand Have you spent much time in Nova Scotia? I'm thinking about submitting to the SQL Saturday there and hanging out for a few days, but wondering if it is worth it. (I realize this is potentially a dumb question because I can't expect Canadians to have visited every part of their country, but I thought I would ask anyway on the chance that you have).
@mmarie i'd like to see some of the shitty sql it throws against servers. you can have all the cool features in the world, but if you generate 50-pages of sql because you have a BA moron dragging and dropping cartesian products all over the place then it doesnt really matter
It's like they tried to go old school for marketing on this release. At their release party they started talking about how they were inspired by "flow". I had an interview question on flow as a high school student trying to get into college. NOw they are rocking all kinds of old memes.
@MarkSinkinson we use cognos so we have idiot devs and BAs using an idiot orm reusing cartesian products because ... well ... just because. but somehow it's the database server's problem.
i'd like to join a consulting group so that when you see the stupidity and make recommendations and they respond with, "GFY" you can say ... "ok ... your money. your time. your problem. NEXT."
Other than that, you suffer through it and realize that you just do the best you can and you don't own the product you are creating and it wasn't how you would do it and the project will be over in 3 months.
"you don't have any way of knowing if a candidate answers truthfully", isn't this valid for every answer?. And I get it, you don't find the question useful, I don't know what more do you want — Lamak20 secs ago
I know there is probably a duplicate floating around the web somewhere, but all my searches couldn't turn up how exactly candidate score is calculated. Is it based on badges? Reputation? Both?
Just passed 40k links to http://sqlfiddle.com on #StackOverflow : http://stackoverflow.com/search?tab=newest&q=url%3a%22http%3a%2f%2f*sqlfiddle.com%2f*%22 !!
I have a self-referencing table Foo
[Id] int NOT NULL,
[ParentId] int NULL, --Foreign key to [Id]
[Type] char(1) NOT NULL
[Id] is clustered primary key, indexes on [ParentId] and [Type] columns.
Assume a maximum depth of 1 on the hierarchy (child nodes cannot have child nodes).
I want to...
I wonder how many people watch some of Adam Machanic's presentations and then go home and bork their server because he is a mad genius and they are not.
I’ve delete and rewritten this question as I don’t think I was being clear and it caused some frustration.
I have a SSIS package that creates and populates an Excel sheet, the package runs perfectly and creates the files as desired and can be executed on a schedule from SQL Server Agent without...
He keeps saying that if he deploys something to prod and it doesn't work that the India dba team should just make it work instead of backing the change out.
They don't know the intent of your process. Stop releasing crap.
The release to production should produce no outcome that was not encountered in pre-production environments. If this is not the case, the change is backed out and you wear a badge of shame
It doesn't matter. Cause you may find it little and suggest a different solution. This is not a what is the right solution type of question. Several information has been left out for simplicity. I can tell you that the current schema needs re-engineering and that the decision has been that the reporting tables will match what we display on the screen which isn't currently the case. For archiving purposes (amongst other reasons) we cannot keep all data in one big table. We need to be flexible in what we can delete without blocking ongoing reporting on those tables. — Yannis53 mins ago