twitter.com/mobileck/status/1131645288014237696?s=19

Morning

Morning

Morning

Morning

@ypercubeᵀᴹ Um, so I can't eat anything but popcorn while you watch a movie? ;)

Morning

@AndriyM right! You all eat popcorn

Fair enough

Good morning

Good morning, database diehards

@Philᵀᴹ I think this q is for you:

You are the Perl expert in the room

@ypercubeᵀᴹ I'm about as good at Perl as Trump is at running Americaland

oh, my memory failed me then. I thought you were using Perl in a previous job.

@Philᵀᴹ Are you a very stable genius too?

@ypercubeᵀᴹ Oh, I have done over the years. I Python now

And have started dabbling with R

This was answer yesterday by @PeterVandivier but the guy accepted another answer...

I will give the answer the benefit of the doubt (about their intentions) and not downvote for the moment.

Feel free to shame me if my comment is wrong

i mean... i was sort of resisting the urge to downvote as well

screw it, it's factually incorrect, misleading, and kind of arsey in tone, downvote incoming

i was actually thinking about the branding problems "Point-in-Time" recovery has as a label for people who aren't already quite familiar with it

like... all times are points in time...

also look at my hands, how weird are my hands

Guys, guys, you might be underappreciating the one aspect to that answer that may be most beneficial to the OP, and that is: it says exactly what the OP wants to hear.

@AndriyM this-guy-gets-it.gif

The question is on the HNQ now :D

that makes me a Saad Butt

Clickety

@PeterVandivier Yes, me too. But actually on your answer. The reason? Well just running Ola's script does not ensure your databases receive the TLC they require. Huh? Yes, Ola's scripts don't come with a schedule. (you won't find any sp_add_jobschedule steps in the script) You need to implement a schedule for the jobs and seeing as the OP is not very fluent in backups, I would say that your answer is missing some points. Sorry if I'm a bit too blunt pointing this out.

@hot2use i had actually forgotten that 😬 should probably update the answer eh?

ty :-)

guess it's been a minute since i ran the raw script from Ola w/o my own customizations

@PeterVandivier I think that the community would benefit from including that tiny bit of information ;-)

And I will remove the comment from your answer once you have added the info.

:p

Well, I just VtC on that question, because it is way too broad IMO. I find it just too generalised to answer.

I mean consider this question: "Is mixing red and green the same as just buying yellow?"

of course not, you get brown ;)

The answer: It depends™

Are there too many scenarios to account for, though?

It depends on the shade of yellow you are after.

I reworded the accepted answer. It is an answer, but lacking a lot of IFS and WHENS.

Well, speaking about the question you VtC'ed, basically you need to cover two cases: 1) you need PITR; 2) you don't need PITR. Is there much more to it?

I mean, that's what need to be stipulated answering the OP's main question:

> My exact question: Is generating multiple daily backup with simple recovery model, similar to full recovery backup with longer period?

@AndriyM "Similar" makes it broad. "Same" would make it answerable. (IMO)

I will let the community decide. :-)

@hot2use Makes sense to me, thanks.

@hot2use editted to try and incorporate the necessary info, but of course feel free to add info/make a community answer if you feel i've missed something important

@PeterVandivier I have removed my comment on your answer. I will give the question a thought or two and might add some information to your answer.

I love backups and restores

:)

e.g. (Ola Hallengren) Deleting log backups older than last FULL backup

ugh ...

How to create a new Apple ID and possibly How to create an anonymous user account using the users Apple-ID like Day-One does?

Nice one

:-)

The link is valid, if you change the i in the appropriate place.

;-)

And if you don't see which i needs fixing, then maybe you need to fix your eye(s) :)

I wonder how you could change the parsing semantics of sql to indicate which rdbms you were using.

-- sql: pgsql
// pgsql

-- sql: mssql;
// sql server

@EvanCarroll SQL_COMPAT global variable

sure, but for all rdbms's

the hinting mechanism in SO is probably exactly what I want I just want it for all-code

I think I'll RFC that.

for that you'll need every DBMS to implement every other DBMS' dialect features. Good luck with that

Looking for advice on file format/configuration system. I have complex transform rules and right now the informaticist can edit an Excel workbook and I can validate this and pull it into configuration tables. But the Excel workbooks need to get more powerful to enable editing in certain ways and lookups and parent-child relationships. In addition, putting these in version control is problematic because you can't easily get usable diffs.

XML would be ideal since it can get diffed pretty effectively and is easy to validate and just as easy to load. But editing tools? Would need something pretty easy and customizable. Considering using Excel (for editing) -> XML (for diff/version cotnrol) -> SQL (runtime) so that the build process maintains the XML automatically.

XML is so 20th century. Go with YAML, you can use any old text editor with it, and it still gives you hierarchical structure and schema validation.

Well, that's just it, I don't want the informaticist to have to know the syntax at all.

are you putting the relationships into excel?

or do you want to provide that external to excel?

Everything the informaticist needs should be in whatever they are editing it.

For example, just for them to be able to review the data modeling with hospital analysts, I already have a dedicated tool which generates sample data from some combinations they can create so they can see if it makes sense in PowerBI before we even worry about populating the model from the source data using the mapping rules.

There's no more syntax in YAML than there is in a bullet list in a Word document. Any informaticist should be able to handle bullet lists. those who can't should bite the bullet.

Are there any tools which would let them have pick lists?

There are going to be thousands of these mappings.

The kind of thing they are doing in the modeling is taking a particular observation in a cardiac catheter study and deciding that it has a severity and then some child attributes. And the child attributes may be reused between different types of observations.

If there is a YAML editor which has some kind of schema validation/intellisense, that might work.

It's hard enough just trying to get the model validated in PowerBI, LOL.

Currently sitting through a meeting where trying to make PowerBI do things that it isn't very good at.

@CadeRoux Fake data?

YEs

+1

A big part of the job is just trying to make representative fake data since real data isn't really ever leaving the hospital.

This meeting really reminds me that no matter what you do, you can never make it easy enough for someone to to do all the analyses that they will think of. Eventually someone will need to write queries.

The whole purpose of our structured reporting system is to get users from having to make copious free text entries like this to structured data that can be analyzed.

@CadeRoux If I were doing it for myself, I'd try to configure this with VS Code code completion and snippets.

Such a mixture of questions on the site at the moment. There’s some really good, hard, SQL Server questions, then a load of utter dogcrap 🐕 from students

@Philᵀᴹ :-)