« first day (2190 days earlier)      last day (2284 days later) » 

12:22 AM
Would you expect a performance difference between N[Re[expr], p] and Re[N[expr, p]]?
 
12:34 AM
@VladimirReshetnikov Can you clarify? I certainly would expect a difference because Re[expr] might involve transformations that are easier if you have evaluated expr numerically before.
 
I thought N[Re[expr]] might see that only real part needs to be evaluated and could save time by not evaluating the imaginary part.
 
@VladimirReshetnikov Do you have an example that you can share?
@VladimirReshetnikov Have you seen this:
In[20]:= Trace[Re[Sin[4/2]],TraceInternal->True]
Out[20]= {{{{1/2,1/2},4/2,2},Sin[2]},Re[Sin[2]],{Sin[2.000000000000000],0.9092974268256817},Sin[2]}
It seems that Re itself is making numerical evaluations on its run.
 
For example, N[Re[PolyLog[3, 2/5 - (2 I)/5]], 5000] seems a little faster than the other way.
 
@VladimirReshetnikov Definitely not for me:
 
I restarted the kernel before each evaluation
 
12:46 AM
That might be an issue due to caching..
 
But the difference seems to be less than 1%, if it exists at all.
 
I still get
@VladimirReshetnikov Your logic, however, seems reasonable if N would indeed takes the Re into account and uses a different/faster algorithm.
 
 
9 hours later…
9:24 AM
@halirutan another illustration of the weird incentives of MSE. A post on using the NASA APIs that took me ~1hr to knock out is now my most upvoted post on SE, while things that took days to develop and some serious design work languish (relatively).
 
 
2 hours later…
11:23 AM
@b3m2a1 Well, is that wrong? Noone has said that that is most useful for the community which takes the longest to write.
 
 
2 hours later…
1:21 PM
@C.E. @b3m2a1 I what b3m means is the other way around: It is sad when you have an answer that is interesting and took you long but this is not reflected in votes, because e.g. the field is very special.
@b3m2a1 One of my top answers took me no time at all because it was only an information, indeed it was a SyntaxInformation (haha, what a word-play). If I had known that this is not commonly known, I would have posted it long before.
Sometimes, I advocate some of my answers here in chat when I think I put a considerable effort into finding the solution, but the question is so specific that it would go unnoticed. E.g. reading the CCompiler code for this answer and solving why it doesn't work took quite some time.
Please note that this question is from 2015 and I answered it almost 3 years later :) I stumbled over it by accident.
 
 
1 hour later…
2:50 PM
Here's an example of a post with a lot of votes that took almost zero effort on my part, mathematica.stackexchange.com/a/115495/9490, but then this one was super fun and took forever, didn't get much attention.
 
3:10 PM
lol - thanks for the upvote :-)
 
3:21 PM
Np. I try and upvote answers that show a significant amount of effort, even if they aren't directly relevant for me.
 
I'm going to leave this here ...
0
Q: Keep only a subgraph of a larger graph while preserving all graph properties

SzabolcsI am looking for a way to delete some edges/vertices from a graph and preserve others, while correctly managing/preserving all graph properties. I came up with the following function: Slice[graph, subGraph] will keep only those edges/vertices in graph that are present in subGraph. An alternati...

One of the simpler questions is: can you break this function?
 
3:56 PM
Curious: What did this symbol "\[InlinePart]" used to do? And which version did it do it?
 
@Edmund My memory is not great, but I think that may have come up during the 10.0 beta versions when the association syntax wasn't yet fleshed out.
 
@Szabolcs Thanks.
@JasonB. Made it a bit simpler - mathematica.stackexchange.com/a/163806/19542
 
5:03 PM
hey chat.mm.se folks
does any of youse have any juicy gossip to share about what kind of timeframe v11.3 and v12 will be released in?
=)
 
5:34 PM
@EmilioPisanty Only WRI knows.
 
@Edmund I actually asked about this ~ a year ago
Sorry two years ago. There's an answer by R. M. on here somewhere.
 
6:13 PM
Given a dataset with the dimensions `id, age, sex, region, income, married, children, car, save_act, current_act, mortgage, pep` - what does `save_act, current_act` and `pep` (resp.) mean? They are boolean variables (yes/no).
I'm not a native speaker and i cant really make sense of it.
 
@ngmir I think that'd be hard even for native speakers without extra context
Usually a dataset will include some form of manifest where they tell you the column labels. What entity did you get that dataset from?
 
It's part of an exercise sheet for an university course. I would ask the person who made the sheet tomorrow but I figured if its just some abbreviation I'm not familiar with I could maybe get a quick answer right now.
 
@ngmir it is not obvious to me, but perhaps someone else will see it
 
:) thanks anyway
 
6:32 PM
@ngmir Has a savings account, current account (how is that different? idk), personal equity plan are my guesses. Finally found someone who includes defs: docs.huihoo.com/weka/… (page 9). Most of it is in Chinese, but not the definitions.
 
@MichaelE2 I was thrown by current_act. No idea what that would be.
 
@b3m2a1 Not clear to me either.
 
6:49 PM
@MichaelE2 ahh, nice! Thank you very much!
 
@ngmir You're welcome! :)
 
 
2 hours later…
8:45 PM
Have there been any improvements to importing big text files. I have a 2GB TSV with each row having 5 numbers, then a one word string (no quotes), and then a number. I cannot use Import but those in R can easily import in under 3 mins.
I have to skip the first two lines get to the data. Trying import crashes the computer as memory goes through the roof
 
@Edmund - good news is there have been improvements in this area
bad news is they are not in the current releases
for now you still need to set it as a stream, and loop through it using Read or ReadLine, or some such
 
9:37 PM
@JasonB. Fingers are crossed.
 
10:26 PM
Does anyone know if Belisarius might come back?
 
 
1 hour later…
11:31 PM
0
Q: Should we consider the vote count when voting?

anderstoodIt has probably been already raised in some meta over SE. The question is, should we vote "relatively", i.e. based on if the question/answer is useful/well-written/etc., or "absolutely", i.e. by assigning a reasonable value to the question and voting in that way. For example, if I see a good que...

 

« first day (2190 days earlier)      last day (2284 days later) »