Krankenstyle posted:

hrmm

i have a list of dicts where a certain key contains a list. i want a merged list of all those lists, but my list comprehension comprehension is lacking... halp?

code:

>>> columns = [{'a': [1,2,3]}, {'a': [4,5,6]}]
>>> [c['a'] for c in columns]
[[1, 2, 3], [4, 5, 6]]
>>> [t for t in c['a'] for c in columns]
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
NameError: name 'c' is not defined

e: doy

code:

>>> [t for c in columns for t in c['a']]
[1, 2, 3, 4, 5, 6]

Fatty Crabcakes posted:

4) brief description, ISSUE-42069

Zlodo posted:

This. It's super useful when a super long time later you modify some code and it does something puzzling that you need to remove or modify and you can look up the jira that this was fixing and verify that your changes are not making the bug reappear

cinci zoo sniper posted:

the fast solution to that, just fyi, is an .expand() for loop

Krankenstyle posted:

you mean .extend()?

Zlodo posted:

This. It's super useful when a super long time later you modify some code and it does something puzzling that you need to remove or modify and you can look up the jira that this was fixing and verify that your changes are not making the bug reappear

we have a perforce commit plugin where we can reference the bug we're fixing so it automatically gets closed and cross referenced with the changelist number, it's super useful

Corla Plankun posted:

if you're commiting a list comprehension that is more complicated than something like

code:

[butt>0 for butt in butts] 

I strongly recommend that you rewrite it as a for loop

if you don't, you're gonna be real mad at yourself later for being so clever

Xarn posted:

That's what the tests are for?

Corla Plankun posted:

if you're commiting a list comprehension that is more complicated than something like

code:

[butt>0 for butt in butts] 

I strongly recommend that you rewrite it as a for loop

if you don't, you're gonna be real mad at yourself later for being so clever

Zlodo posted:

We do little to no automated testing unfortunately. And what we do is very high level testing, things like automatically running AIs in the various racing events of our game and see if they get stuck.

Our qa team have a smoke test procedure to test the major features of the game every time they get a new build, but it's easy to break something that no one will notice - the combinatorial complexity of our features and game flows and game states and the data that it all feeds from is just too much to cover everything.

Some unit testing might help, but the humongous code base was absolutely not designed with unit testing in mind. And then there's the fact that most of our bugs are caused not by individual modules but by the interactions between multiple modules and what the player have in their save and what designers have setup in the data.

DONT THREAD ON ME posted:

well, that sounds like a nightmare

Fatty Crabcakes posted:

Because you touch yourself at night.

Powerful Two-Hander posted:

what the gently caress why does this method pass an instance of a class to itself as a parameter for its own save method

Powerful Two-Hander posted:

what the gently caress why does this method pass an instance of a class to itself as a parameter for its own save method

Powerful Two-Hander posted:

what the gently caress why does this method pass an instance of a class to itself as a parameter for its own save method

Blinkz0rz posted:

if i had to guess someone probably wrote a dal abstraction layer that uses the class type as an entity to pass into a factory that saves the right object type but then forgot what they were doing and gently caress you that's why

cinci zoo sniper posted:

yeah

Krankenstyle posted:

weird that list comprehensions are slower than for-loops. youd think theyd be at least the same speed

Krankenstyle posted:

weird that list comprehensions are slower than for-loops. youd think theyd be at least the same speed

Blinkz0rz posted:

dal abstraction layer

Achmed Jones posted:

oh so there’s this rule of software development that also holds true for devops, sre, security engineering, and occasionally life. the fundamental law is: because gently caress you, that’s why

Private Speech posted:

a dalal if you will

Blinkz0rz posted:

cinci zoo sniper posted:

i meant to say that this is definitely faster than itertools.chain, though i wouldn’t be surprised if in general case extend still is the fastest - im not familiar with its implementation

galenanorth posted:

I'm running the last CSV-creating program, for all the Fifth Third Bank branch locations and ATMs, before I register as an official business. It's got 3 days to go until it finishes running because the FCC's API for adding county data given lat-long data is slow and I've been putting off learning to do client-side GIS with Python

In the meanwhile I'm getting started on scraping data from Kroger, a grocery store chain which has its data available as a JSON object buried in the HTML source code of its store locator page. After I list my business on Google Ads for a month, if it doesn't bring in enough money, I'll get a full-time job and switch to working on the project on weekends. My parents have been sick and my mom is getting colon surgery soon. The project is the only reason I haven't gotten a job yet, but I can always quit the job if I need to in order to take care of my mom when she gets back from surgery.

cinci zoo sniper posted:

i meant to say that this is definitely faster than itertools.chain, though i wouldn’t be surprised if in general case extend still is the fastest - im not familiar with its implementation

Symbolic Butt posted:

is it? I mean, it is if you're converting the iterable to a list sure but... well, if you're working with lists you shouldn't be worrying about speed really tbh

DONT THREAD ON ME posted:

lol at the very idea of having to care about iterator performance in a plang

perhaps if performance mattered to you you should choose a language with zero cost abstractions, like rust

gonadic io posted:

i'm interviewing at a company that claims to be doing HFT ...in f#

DONT THREAD ON ME posted:

coming back to a managed langs after working on unmanaged is weird.