The Heap™ – Consultancy ©® - 2016-12-09 (page 1 of 2)

6:49 AM

I'm voting to close this question as off-topic because it is a cross-post. The other instance of the question already has an answer at the time of writing this. Please do not cross-post. — Andriy M 2 mins ago

Andriy M

7:32 AM

The other instance has been deleted, vote withdrawn.

hot2use

8:09 AM

Good morning, and in case I don't see you, good afternoon, good evening, and good night

Andriy M

Good morning

James Anderson

8:27 AM

Does anyone here have any views on Entity Framework? Struggling to see the benefits.

hot2use

8:44 AM

@JamesAnderson No views. Never used it, but I'm not THE programmer/developer guy these days

Andriy M

Easy class/table coupling in an application. Easier data layer handling when you are not very good at SQL. All the benefits I can see are mainly (client-side) development-related. But I've never used it either, my conclusions are based on encounters in SE.

Tom V

@JamesAnderson depends on the size of database in my opinion

A group in our company uses it for a piece of software, it seems to work fine for smaller installs, but it's a bit hard to keep performance under control on larger setups

But that might just be the way they are using it and not inherent to EF, I don't know

I don't have extensive experience on the database side, and none on the development side

They also keep talking about code first and database first, I have no idea what the implications are, but you might want to look into that

James Anderson

@TomV Yes I've heard these terms

I think they are saying that you define a class in the 'model' and then the SQL to create the table is generated.

Tom V

One thing to make sure is to watch out for inheritance, it looks nice at first, but you can't create a composite index across tables of course

James Anderson

I thought that sounded quite nice because any conflicts when merging code branches would highlight schema change conflicts but seeing the work required to add 1 column I'm not sure it's worth it

@TomV interesting

ypercubeᵀᴹ

9:17 AM

@JamesAnderson We use sqlalchemy like that. It's sometimes a pain.

As I have to make sure after each change in the model that the SQL generated is what I wanted

The schema changes are handled by alembic actually (another python module)

It's not too bad when you get used to it.

And I have no idea how EF handles all these (same or differently)

James Anderson

9:33 AM

Yesterday I was asked to look at why an EF generated query was slow.

After running a trace I could see that the query was joining two tables and then doing a SELECT TOP 1 * FROM ... WHERE a = 'xxx' from a third table for each row

Jack Douglas

begin;
create schema stack;
set search_path=stack;
--
create table foo as
select *
from (values (1,'X','D','G','P')
, (2,'F','D','L','M')
, (3,'X','N','R','S')
, (4,'Y','I','W','Q')
, (5,'U','Z','E','H') ) AS f(id,a,b,c,d);
--
with recursive t(id,a,b,c,d,start,path,cycle) as (
select *, id, array[id], false from foo
union all
select f.*, start, path||f.id, f.id=any(path) from foo f join t on f.id<>t.id and (f.a=t.a or f.b=t.b or f.c=t.c or f.d=t.d) where not cycle )
select array_agg(f.id order by f.id) ids

ypercubeᵀᴹ

@JackDouglas Nice.

Jack Douglas

it scales extremely badly :)

James Anderson

After playing about a bit we found that it does this because there is no FK releationship between either of the first 2 tables and the 3rd. There is a way to force a join without a FK but it got very messy

Jack Douglas

@ypercubeᵀᴹ did you go to the dba drinks last night?

ypercubeᵀᴹ

9:38 AM

@JackDouglas yes. but arrived more than fashionably late ;)

Jack Douglas

what, like just before chucking out time?

ypercubeᵀᴹ

about an hour, or a bit more

dezso

@JackDouglas why does the reference arrow not work? Or is it only me?

Jack Douglas

@dezso are you ignoring anyone who frequents the room?

it takes me here:

11 hours ago, by Evan Carroll

I really don't want it answered there, but here is the teaser. http://stackoverflow.com/questions/41045859/group-by-any-of-multiple-columns

dezso

@JackDouglas but of course

I see, that's the explanation then

Jack Douglas

9:45 AM

interesting, I'll have to file that in my brain somewhere

Philᵀᴹ

@dezso This came from that: dba.stackexchange.com/questions/157639/…

dezso

@Philᵀᴹ thanks

Andriy M

But the SO question is where it started. Evan is probably researching options for a potential answer.

Jack Douglas

10:37 AM

@AndriyM He is going up a blind alley with CUBE though — I don't think any solution is possible that doesn't use recursion.

because two rows can end up in the same group even though they have no data in common

hot2use

1:05 PM

I'm glad it's Friday :-)

ypercubeᵀᴹ

2:22 PM

@ThomasWard you around?

Thomas Ward

@ypercubeᵀᴹ what's up?

ypercubeᵀᴹ

odd question: Possible reasons for apt-get install upgrade to need 30+ minutes to run?

Thomas Ward

@ypercubeᵀᴹ slow network connection for downloading the upgrades, slow to compile kernel modules, grub taking a long time to search for kernels to add to the boot options list, any of a number of other things

possible DNS issues too

ypercubeᵀᴹ

The dev had never run it before? Slow network? Low-spec hardware? All of the above?

Ah, thnx!

Thomas Ward

slow network, low-spec hardware will slow down some things, never been run, without a lot more info I couldn't say specifically

ypercubeᵀᴹ

2:25 PM

No worries. It's not too important, to spend time investigating.

Thomas Ward

@ypercubeᵀᴹ yeah, that sounds about right. In this case, I'd say that it could be any of a number of different reasons from a slow mirror to never being run before, to slow system, to ...

(the list goes on)

ypercubeᵀᴹ

It's nice that we have mods from SuperUser, AskUbuntu and SE employees. We don't have to wander the sites to get our questions answered.
The Heap™ knows. Everything. ;)

5

Thomas Ward

@ypercubeᵀᴹ heh

Transcript for

The Heap™ – Consultancy ©®