That's not quite right: the highest performance case requires that you have
three separate functions:
(1) f, which only calculates the objective function as a pure function
returning a Float64 value
(2) g! which only calculates the gradient by mutating a Float64 array
(3) fg!, which calculates
This was a really nice read, even you can tell the author’s spent time with
social scientists based on their assertion that a correlation of 0.3 is low.
— John
On Aug 13, 2014, at 6:41 PM, Samuel Jenkins svs14.41...@gmail.com wrote:
Since this is on the topic of package popularity, this blog
Here are some things you might try:
(1) Don’t use a global const. Pass N as an argument to all functions.
(2) Use in-place multiplication via functions like A_mul_B! in lines like
`@inbounds a = dot(W[:,n], state)`. In Julia 0.3, W[:,n] creates a copy of that
column of W on every iteration.
Not possible in the current versions of Julia. Maybe one day. There are bunch
of us who’d like to have this functionality, but it’s a non-trivial addition of
complexity to the compiler.
— John
On Aug 14, 2014, at 4:59 AM, Chris Kellendonk chriskellend...@gmail.com wrote:
I'm new to the
In Julia, anything with a description like 1x27 Array{Any,2} is a matrix, not a
vector.
Also, it's not the right kind of vector because it has no numeric type
restriction. It needs to be a vector of Float64 values, like you'd get from
doing something like:
[1.0, 2.0, 3.0]
-- John
On Aug
, 1}?
Thanks again!
Pam.
On Thursday, August 14, 2014 5:15:01 PM UTC+1, John Myles White wrote:
In Julia, anything with a description like 1x27 Array{Any,2} is a matrix, not
a vector.
Also, it's not the right kind of vector because it has no numeric type
restriction. It needs
I just heard a talk about this (somewhat out-of-date) LLVM plugin for doing
performance testing in a way that randomizes over possible memory layouts:
https://github.com/ccurtsinger/stabilizer
If the plugin gets updated for a newer version of LLVM it might be a fun
project to try using this
I think Lint.jl might handle this already: https://github.com/tonyhffong/Lint.jl
-- John
On Aug 14, 2014, at 2:39 PM, Ethan Anderes ethanande...@gmail.com wrote:
It's funny, this issue is by far my biggest source of bugs when coding in
Julia. I find myself often prototyping things with
a . 5
On Aug 15, 2014, at 3:53 AM, David Higgins daithiohuig...@gmail.com wrote:
Hi,
Is there a mechanism for vectorised logical access to array elements in
Julia? I basically mean, is there an equivalent to the Matlab notation
a[a5]
which should give all of the elements of a[] which
I agree: I think most people should not use 0.4 beause it’s going to change so
frequently.
— John
On Aug 15, 2014, at 12:28 AM, Tobias Knopp tobias.kn...@googlemail.com wrote:
I think we should discourage using 0.4-dev version until feature freeze.
There might be a lot of package breakage
Hi Brendan,
It looks like you’re hitting Julia’s invariance for the first time:
http://en.wikipedia.org/wiki/Covariance_and_contravariance_(computer_science)
If you’re working with inputs to a function, you can do the typecheck using a
type parameter, T:
function foo{T :
, you get
invariance.
— John
On Aug 15, 2014, at 7:58 PM, Brendan O'Connor breno...@cs.umass.edu wrote:
On Fri, Aug 15, 2014 at 9:53 PM, John Myles White johnmyleswh...@gmail.com
wrote:
Hi Brendan,
It looks like you’re hitting Julia’s invariance for the first time:
http
Try this section of the manual:
http://julia.readthedocs.org/en/latest/manual/networking-and-streams/
— John
On Aug 16, 2014, at 6:52 PM, frbertonce...@gmail.com wrote:
Hi all,
just start to look at this julia but it seems there is not syntax for I/O
data..
Does anyone know if julia
This is great. Thanks for sharing this.
— John
On Aug 17, 2014, at 2:35 PM, Don MacMillen don.macmil...@gmail.com wrote:
I've made a Julia version that is available here http://bit.ly/1pi53Ih
I hope you find it useful.
I can probably register KNN.jl and deprecate kNN.jl.
— John
On Aug 18, 2014, at 5:45 AM, Iain Dunning iaindunn...@gmail.com wrote:
Yes, it is the Julia convention.
There is only one (registered) exception, which is prehistoric in Julia terms
(kNN.jl)
On Monday, August 18, 2014 8:32:40
systems like HFS+ on Macs.
On Aug 18, 2014, at 10:08 AM, John Myles White johnmyleswh...@gmail.com
wrote:
I can probably register KNN.jl and deprecate kNN.jl.
— John
On Aug 18, 2014, at 5:45 AM, Iain Dunning iaindunn...@gmail.com wrote:
Yes, it is the Julia convention.
There is only
Hi Bradley,
Would you consider using DataArrays for this? DataFrames no longer support
these operations, so any upgrade in your setup would turn all of this code into
errors.
All of these operations work on DataArrays already.
— John
On Aug 18, 2014, at 7:28 AM, Bradley Setzler
, 2014 9:31:43 AM UTC-5, John Myles White wrote:
Hi Bradley,
Would you consider using DataArrays for this? DataFrames no longer support
these operations, so any upgrade in your setup would turn all of this code
into errors.
All of these operations work on DataArrays already.
— John
One of the best things you can do for the Julia community is to start using the
release candidates of 0.3. Everyone will be better off if bugs are found now
rather than when 0.3 is released. You can help make 0.3 a better release by
installing one of the latest binaries and testing it out.
not convert if divided
by Integer.
Best,
Bradley
On Monday, August 18, 2014 9:31:43 AM UTC-5, John Myles White wrote:
Hi Bradley,
Would you consider using DataArrays for this? DataFrames no longer support
these operations, so any upgrade in your setup would turn all of this code
I don't think this is possible. This sounds like a case where PDMats needs a
new release. You can implement a release yourself, make a pull request to
METADATA and ask Dahua for permission to merge it.
-- John
On Aug 19, 2014, at 10:13 AM, Spencer Lyon spencerly...@gmail.com wrote:
Is there
We should really have a FAQ for this. This is something that tripped me up for
a while.
I’ll write up sometime.
— John
On Aug 19, 2014, at 9:11 PM, Isaiah Norton isaiah.nor...@gmail.com wrote:
That is the splice operator (also known as splat). It causes the contents
of `a` to be passed as
Copying things by default isn’t a very good strategy for a language used to do
linear algebra on large arrays.
Requiring that people use Ref to get reasonable performance for linear algebra
operations would make Julia feel like a much more low-level language than it is
currenty.
— John
On
? It looks like there is some debugging code in
there but I'm not sure how to turn it on.
-thom
On Wednesday, July 30, 2014 6:24:26 PM UTC-5, John Myles White wrote:
I’ve never seen our line search methods produce an error that wasn’t caused
by errors in the gradient. The line search methods
search routines? It looks like there is some debugging code in
there but I'm not sure how to turn it on.
-thom
On Wednesday, July 30, 2014 6:24:26 PM UTC-5, John Myles White wrote:
I’ve never seen our line search methods produce an error that wasn’t caused
by errors in the gradient
Try typing the sequence \_1TAB
— John
On Aug 20, 2014, at 8:08 AM, Douglas Bates dmba...@gmail.com wrote:
I find it convenient to use identifiers like
julia a₁ = 2
2
In the iESS emacs mode for a Julia session I can set the input-mode to TeX
and type a_1 to create this identifier.
Unfortunately, the rule is simple: dispatch never looks at keyword arguments.
-- John
On Aug 21, 2014, at 8:59 AM, Spencer Lyon spencerly...@gmail.com wrote:
Consider the following code:
function periodogram(x::Array)
n = length(x)
I_w = abs(fft(x)).^2 ./ n
w = 2pi * [0:n-1]
Abstractly, I’d love to be able to print using Color.jl. It seems like a much
cleaner interface.
But it’s not clear to me that Color.jl buys you a lot if your terminal doesn’t
support a widerange of colors.
— John
On Aug 22, 2014, at 3:33 PM, Spencer Russell spencer.f.russ...@gmail.com
a = rand(5000,5000); b = rand(5000); x =a\b end;
elapsed time: 30.509950844 seconds (400100776 bytes allocated, 0.11% gc time)
It's mystery to me. If I learn anything more, will let you know.
Thanks again.
Don
On Friday, August 22, 2014 9:54:22 PM UTC-7, John Myles White wrote:
I don’t
Can you clarify a bit what you mean by “same result”?
It would also be great to see a simpler example that doesn’t involve the
Bayesian updating steps, which hopefully aren’t essential to hitting an error.
— John
On Aug 23, 2014, at 4:33 PM, asim yahooans...@gmail.com wrote:
I am trying
:
That is no longer valid under in Julia 0.3. The code only runs 3x faster and
the type stable code looks a bit more complex.
On Monday, December 16, 2013 8:48:40 PM UTC+2, John Myles White wrote:
I think Michael is suggesting that it would help to have an automated tool to
do this. This stuff
looks a bit more complex.
On Monday, December 16, 2013 8:48:40 PM UTC+2, John Myles White wrote:
I think Michael is suggesting that it would help to have an automated tool to
do this. This stuff comes up in subtle ways in complex code. While editing
some of the t-SNE code someone posted
The NEWS.md file does cover this:
• Ranges and arrays with the same elements are now unequal. This allows hashing
and comparing ranges to be faster. (#5778)
On Aug 25, 2014, at 8:45 AM, Ed Scheinerman edward.scheiner...@gmail.com
wrote:
Thanks again for the pointer to the release notes.
The issue is that you want to have all code documentation show up in REPL. In
the GoDoc approach, this might require an explicit build step -- which is a
non-trivial cost in usability.
-- John
On Aug 25, 2014, at 3:01 PM, Job van der Zwan j.l.vanderz...@gmail.com wrote:
On Monday, 25 August
This looks like a failure to find functions from NumericFuns.
What versions of Julia and stats packages are you using?
-- John
On Aug 25, 2014, at 9:03 AM, asim yahooans...@gmail.com wrote:
mvCanonNotebook.pdf
Dahua, did you keep the original package around?
— John
On Aug 25, 2014, at 10:07 PM, Dahua Lin linda...@gmail.com wrote:
Following Julia package naming convention, the package Distance was renamed
to Distances.
New package page: https://github.com/JuliaStats/Distances.jl
All
Dahua
On Tuesday, August 26, 2014 1:08:48 PM UTC+8, John Myles White wrote:
Dahua, did you keep the original package around?
— John
On Aug 25, 2014, at 10:07 PM, Dahua Lin lind...@gmail.com wrote:
Following Julia package naming convention, the package Distance was renamed
...@gmail.com wrote:
On Tuesday, 26 August 2014 00:04:41 UTC+2, John Myles White wrote:
The issue is that you want to have all code documentation show up in REPL. In
the GoDoc approach, this might require an explicit build step -- which is a
non-trivial cost in usability.
-- John
I assume you
1.0
1.5 2.0
julia A/2.
2x2 DataArray{Float64,2}:
0.5 1.0
1.5 2.0
julia A/2
InexactError()
So it converts to Float if divided by Float, but does not convert if divided
by Integer.
Best,
Bradley
On Monday, August 18, 2014 9:31:43 AM UTC-5, John Myles White wrote:
Hi Bradley
Hi Phillip,
Could you provide a more complete example of what you're doing? In particular,
I'm very confused by the idea of pushing items to a dict, which is an
unordered data structure.
Here is how dictionaries work in Julia:
d = Dict()
d[foo] = bar
sizehint(d, 100)
d[bar] = foo
-- John
sum(A, 1) works, as does sum(A, (1, )).
More generally, sum(A, dims::Integer...) works, as does as sum(A,
(dims::Integer...), ).
-- John
On Aug 27, 2014, at 2:44 PM, Andrew Dabrowski unhandya...@gmail.com wrote:
In the doc for the Standard library I see:
sum(A, dims)
Sum elements of an
as a separate pass. That doesn't strike me as completely out of
the question, but wouldn't be optimal.
On Tue, Aug 26, 2014 at 11:32 AM, John Myles White johnmyl...@gmail.com
wrote:
No, I was talking about what I understood to be a design principle of GoDoc:
doc generation and parsing occurs
formats.
On Wednesday, August 27, 2014, John Myles White johnmyleswh...@gmail.com
wrote:
Ok, thanks for clarifying. I also like the idea of strategically placed
comments as automatic documentation.
-- John
On Aug 27, 2014, at 2:54 PM, Job van der Zwan j.l.vanderz...@gmail.com
wrote
! function when generating multiple draws from the
distribution, and this function is not found. I am not sure where this
function is, either.
Thanks
Asim
On Monday, August 25, 2014 6:14:40 PM UTC-4, John Myles White wrote:
This looks like a failure to find functions from NumericFuns
As we're starting to get better ideas for a documentation system, two questions
I have are how we do two things:
(1) Handle documentation of generic functions and their specialized methods
without requiring documentation of all specialized methods.
(2) Handle documentation of functions that
In theory, the compiler could potentially pull the bounds check out of the loop
since the loop parameters are run-time constants, which means that the bounds
check doesn't actually need to happen on each access.
-- John
On Aug 29, 2014, at 1:21 PM, Ed Scheinerman edward.scheiner...@gmail.com
DataArrays has a cut function.
-- John
On Aug 29, 2014, at 11:17 AM, Florian Oswald florian.osw...@gmail.com wrote:
hi
what's the julia equivalent of this R call? i don't need the levels and
labels, just some kind of grouping index.
cut(sample(1:10,10,TRUE),c(0,3,6,10))
[1] (3,6]
be situations where several
packages/functions do the same thing. But it's hard for the user to find find
simple things like that I think. let me know what you think.
On 29 August 2014 22:29, John Myles White johnmyleswh...@gmail.com wrote:
DataArrays has a cut function.
-- John
On Aug 29
Hi Asim,
It’s a little hard to work with PDF’s. Would you consider using Gists?
(https://gist.github.com)
— John
On Aug 30, 2014, at 1:47 PM, asim yahooans...@gmail.com wrote:
Hi
The Blas trsv function is described as needing 7 arguments in the
documentation. However, it only appears
I can’t speak for others, but I’m very hesitant to download any kind of files
from mailing lists.
— John
On Aug 30, 2014, at 2:00 PM, asim yahooans...@gmail.com wrote:
Would this help?
Asim
On Saturday, August 30, 2014 4:49:23 PM UTC-4, John Myles White wrote:
Hi Asim,
It’s
This might need to be part of the Zen of Julia.
— John
On Aug 30, 2014, at 2:11 PM, Jameson Nash vtjn...@gmail.com wrote:
calling eval in a macro doesn't do what you think it does, so it doesn't do
what you want
On Sat, Aug 30, 2014 at 5:05 PM, Don MacMillen don.macmil...@gmail.com
I don’t think this example had any views. Both bindings had an equal right to
be considered the true binding.
I think we’re better off doing more education to teach people to distinguish
bindings and values.
— John
On Aug 31, 2014, at 11:27 AM, Ethan Anderes ethanande...@gmail.com wrote:
Bradley, it’s especially easy to edit documentation because you can make a Pull
Request right from the website.
— John
On Aug 31, 2014, at 11:30 AM, Bradley Setzler bradley.setz...@gmail.com wrote:
Thank you Adam, this works.
Let me suggest that this information be included in the GLM
Merged. Thanks, Bradley.
— John
On Aug 31, 2014, at 12:29 PM, Bradley Setzler bradley.setz...@gmail.com wrote:
Thank you for suggesting this, John.
https://github.com/JuliaStats/GLM.jl/pull/90
Bradley
On Sunday, August 31, 2014 1:33:04 PM UTC-5, John Myles White wrote:
Bradley
something like a series option
for non-parametric regression,
glm(Y,X,data,family,link,seriesRank=2)
where seriesRank=2 means all of the terms X1.^2, X1.*X2, X1.*X3,...,X5.^2 are
included as regressors.
Bradley
On Sunday, August 31, 2014 2:32:30 PM UTC-5, John Myles White wrote:
Merged
I think there’s a broad issue that need resolution: how do you know when a
function’s output takes control of the memory used by its arguments?
— John
On Aug 31, 2014, at 11:45 AM, Ethan Anderes ethanande...@gmail.com wrote:
Yeah, I can see your point John. It's probably not reasonable to
.^2
X2.^2
X3.^2
X1.*X2
X1.*X3
X2.*X3
X1.*X2.*X3
Bradley
On Sunday, August 31, 2014 2:55:22 PM UTC-5, John Myles White wrote:
Bradley, you’re forgetting about interactions terms.
— John
On Aug 31, 2014, at 12:53 PM, Bradley Setzler bradley...@gmail.com wrote:
No problem
be,
(intercept)
X1.^2
X2.^2
X3.^2
X1.*X2
X1.*X3
X2.*X3
X1.*X2.*X3
Bradley
On Sunday, August 31, 2014 2:55:22 PM UTC-5, John Myles White wrote:
Bradley, you’re forgetting about interactions terms.
— John
On Aug 31, 2014, at 12:53 PM, Bradley Setzler bradley...@gmail.com wrote
Have you tried macroexpand?
— John
On Aug 31, 2014, at 10:28 PM, Mykel Kochenderfer m.kochender...@gmail.com
wrote:
I want to do a calculation like this $\max_{a \in A} \sum_{s \in S} g(s, a)$.
Of course, I can do something like this:
maximum([sum([g(s, a) for s in S]) for a in A])
On Sunday, August 31, 2014 12:58:43 PM UTC-7, John Myles White wrote:
I think there’s a broad issue that need resolution: how do you know when a
function’s output takes control of the memory used by its arguments?
— John
Use vec
On Sep 3, 2014, at 11:35 AM, David Smith david.sm...@gmail.com wrote:
Ok, so you can continue using the old squeeze. Us reckless types can use the
aggressive one. ;-)
I don't see why it shouldn't be available. Is there a fear that new users
will run into subtle errors and hate
In general, type inference doesn’t always converge in the global scope.
In this case, you’re seeing that type inference is missing information because
of the reference to cols, which isn’t a compile-time constant.
I personally recommend using typed comprehensions everywhere for explicitness
It is definitely possible. You just need to use toplevel, which is a magic that
Jeff doesn't want people to know about.
-- John
On Sep 5, 2014, at 12:52 PM, Ben Arthur bjarthu...@gmail.com wrote:
sorry to resurrect this old post, but what is the definitive answer to
whether it's possible
How do you feel about using Set( [(1, 2)] )?
— John
On Sep 5, 2014, at 5:01 PM, Sam L sam.len...@gmail.com wrote:
I can't figure out how to make a set of tuples in a very clean way.
This works:
julia push!(Set{(Int, Int)}(), (1,2))
Set{(Int64,Int64)}({(1,2)})
but this doesn't:
I am hoping that the 0.4 release of Julia will coincide with a major cleanup of
the Data* world. I wrote up a very high level overview of my goals here:
https://gist.github.com/johnmyleswhite/ad5305ecaa9de01e317e
There’s still more work to do to flesh out these ideas, but the basic
principles
, John Myles White wrote:
I am hoping that the 0.4 release of Julia will coincide with a major cleanup
of the Data* world. I wrote up a very high level overview of my goals here:
https://gist.github.com/johnmyleswhite/ad5305ecaa9de01e317e
There’s still more work to do to flesh out these ideas
No, DataFrames are not indexed. For now, you’d need to build a wrapper that
indexes a DataFrame to get that kind of functionality.
— John
On Sep 7, 2014, at 9:53 AM, Steven Sagaert steven.saga...@gmail.com wrote:
Hi,
I was wondering if searching in a dataframe is indexed (in the DB sense,
It gets expanded in chunks.
— John
On Sep 7, 2014, at 10:15 AM, Steven Sagaert steven.saga...@gmail.com wrote:
When you start with an empty array and grow it one element at a time with
push!, does the underlying array memory block get copied expanded by one or
in larger chunks (like
:
https://gist.github.com/johnmyleswhite/ad5305ecaa9de01e317e
On Sun, Sep 7, 2014 at 12:54 PM, John Myles White johnmyleswh...@gmail.com
wrote:
No, DataFrames are not indexed. For now, you’d need to build a wrapper that
indexes a DataFrame to get that kind of functionality.
— John
I think we’re still not really interested in promoting the use of NaN as a
surrogate for NULL, especially given that Nullable is going to be added to Base
in 0.4.
Your functions would perform substantially better if you iterated over the
values of A. For example,
function nanmean(A::Array)
.
On Sunday, September 7, 2014 10:32:53 AM UTC-7, John Myles White wrote:
FWIW, I think it’s much easier to index structures if every row has an atomic
existence that is independent of the table it is currently part of. (This is
a big part of my interest in moving away from matrix semantics
whenever it grows beyond it's underlying size. (This might change
when the array gets large, but I haven't looked at that code recently, so I
forget the details.)
Cheers,
Kevin
On Sunday, September 7, 2014, John Myles White johnmyleswh...@gmail.com
wrote:
It gets expanded in chunks
is apparently not the
right way of thinking.
n=5
x_foo=zeros(n,n)
x_bar=zeros(n,n)
for i in [foo,bar]
x_$i = x_$i + n
println(x_$i)
end
x_foo
On Sunday, September 7, 2014 3:07:10 PM UTC-7, John Myles White wrote:
Hi Alex,
You can’t use things like x_$i
I kind of suspect my team (which is the team that invented Hive) isn't likely
to stop using Hive anytime soon.
-- John
On Sep 7, 2014, at 4:50 PM, Steven Sagaert steven.saga...@gmail.com wrote:
On Monday, September 8, 2014 1:37:50 AM UTC+2, John Myles White wrote:
Well, you can write
I think the types of your inputs might be wrong. In particular, I think the
type of the inner containers isn't right: you're getting Array{Array{T,N},1}
when you wanted Array{Array{E,1},1} for some specific E.
-- John
On Sep 7, 2014, at 5:50 PM, Zenna Tavares zennatava...@gmail.com wrote:
I suspect the only way to do this is to change Interact so that it exposes a
minimum time threshold before it registers a state change.
— John
On Sep 8, 2014, at 4:16 AM, Andrei Berceanu andreiberce...@gmail.com wrote:
I have some code along the lines of
f = figure()
@manipulate for
I don't believe this exists, but it should be easy to write by combining ideas
from Julia's cumsum and foldl. You'd end up with cumfoldl and cumfoldr.
-- John
On Sep 9, 2014, at 7:29 AM, John Drummond john...@gmail.com wrote:
What's the idiomatic way of writing Mathematica's Foldlist in
Try vcat/hcat.
-- John
On Sep 9, 2014, at 3:31 PM, Diego Tapias dandrove...@gmail.com wrote:
Thanks for answering!, but what if I want to form a matrix of dimension 2 and
not an array of dimension 1.
2014-09-09 17:27 GMT-05:00 Stefan Karpinski ste...@karpinski.org:
append!(v,w) – it
You want show, not print.
-- John
On Sep 9, 2014, at 3:32 PM, muraveill murave...@gmail.com wrote:
This is really confusing. In the REPL one can evaluate a stream to get useful
info on what the object is:
julia f
IOStream(file test.txt)
But in a script, just f would not work, so for
This sure looks like you're not making any copies when you seem to want copies.
In particular, this line:
cellList[i] = Cell(i, oVector)
probably needs to be
cellList[i] = Cell(i, copy(oVector))
-- John
On Sep 10, 2014, at 6:41 PM, Andre Bieler andre.biele...@gmail.com wrote:
can
some easily parallelizable operations.
-viral
On Sunday, September 7, 2014 11:47:44 AM UTC+5:30, John Myles White wrote:
Yeah, that’s a way more ambitious project. That’ll take at least a year to
make any progress at all. Before I could even begin, I need to finish DBI and
then build up
For future reference, I'd really appreciate you not referring to my writing as
quite erroneous when the claims made are in fact correct.
-- John
On Sep 11, 2014, at 2:38 PM, Mohammed El-Beltagy mohammed.elbelt...@gmail.com
wrote:
In a recent blog by John Myles White
http
This article is really good. Thanks for pointing it out.
-- John
On Sep 11, 2014, at 4:10 PM, asim Ansari asimansari...@gmail.com wrote:
Hi
This article by Duncan Temple Lang on Compiling R could be of interest.
Mentions Juila a few times.
Asim
http://arxiv.org/pdf/1409.3144.pdf
I'd love it if Duncan were to start working on Julia. His ideas would be
invaluable.
-- John
On Sep 11, 2014, at 6:49 PM, Ethan Anderes ethanande...@gmail.com wrote:
For the past year I've been trying to get Duncan to jump to our side of the
tracks and join in on Julia development. I
One point worth making: it's often much harder to estimate the tails of a
distribution than the mean.
The most natural example of this is estimating the mean of a normal
distribution. From theory we know that the population median and the population
mean are the same quantity, so we could use
As Lint keeps improving, I feel like we should probably highlight it more
aggressively in the manual.
-- John
On Sep 12, 2014, at 10:56 AM, Tony Fong tony.hf.f...@gmail.com wrote:
Lint v0.1.2 starts to track variables' type locally within a function
declaration. So your case will correctly
We really need to standardize on a single type that reflects a single row of a
tabular data structure that gets used both by DBI and by DataFrames.
DataFrameRow is really nice because it's a zero-copy operation for DataFrames,
but we can't provide zero-copy semantics when pulling rows out of a
convert
methods than to re-represent rows as OrderedDicts (or whatever datatype).
On Friday, September 12, 2014 12:19:47 PM UTC-5, John Myles White wrote:
We really need to standardize on a single type that reflects a single row of
a tabular data structure that gets used both by DBI
intermediate step).
On Friday, September 12, 2014 3:26:47 PM UTC-5, John Myles White wrote:
I'm not sure that losing zero copy semantics is actually a big performance
hit in most pipelines.
I think much more important is that you can't write generic code right now
because the abstractions
What does that mean? A DataFrameRow can't be easily created without reference
to an existing DataFrame, so this seems like it's either a mechanism for
transferring rows from one DataFrame to another very slowly or a mechanism for
inserting duplicate rows.
-- John
On Sep 12, 2014, at 3:37 PM,
No worries. I'm glad you enjoy my blog posts.
-- John
On Sep 13, 2014, at 11:32 AM, Mohammed El-Beltagy
mohammed.elbelt...@gmail.com wrote:
On Friday, September 12, 2014 8:33:03 AM UTC+3, John Myles White wrote:
For future reference, I'd really appreciate you not referring to my writing
I’m not so sure that we should follow the lead of Octave and R here. Neither of
those languages reify distributions as types, so changes to their RNG’s don’t
affect other operations on those same distributions.
In contrast, the proposed change here would break a lot of other code in
the popular one! On the other hand,
Scipy returns an error).
Regards,
Jamie
On Monday, 15 September 2014 16:03:32 UTC+1, John Myles White wrote:
I’m not so sure that we should follow the lead of Octave and R here. Neither
of those languages reify distributions as types, so changes
).
Regards,
Jamie
On Monday, 15 September 2014 16:03:32 UTC+1, John Myles White wrote:
I’m not so sure that we should follow the lead of Octave and R here. Neither
of those languages reify distributions as types, so changes to their RNG’s
don’t affect other operations on those same
Hi Josimar,
Julia arrays are not indexed using parentheses: they are indexed using brackets.
Assuming you have an array like x = zeros(10), you can do
for i in 1:10
x[i] = i*i
end
You also don’t need semicolons in Julia.
— John
On Sep 15, 2014, at 5:15 PM, Josimar Silva
You are a genuine Julia community hero, Tony.
-- John
On Sep 16, 2014, at 9:17 AM, Tony Fong tony.hf.f...@gmail.com wrote:
Thanks to this discussion, I have added some performance related lint
messages related to type declaration. It would catch dimension-less array
field type going
Hi Nils,
Try something like:
A = Array(Any, 10)
for i in 1:10
A[i] = randn(1, 10)
end
On Sep 18, 2014, at 6:47 AM, nils.gu...@gmail.com wrote:
I'm sure this is an extremely trivial question, but I can't seem to find an
answer anywhere. I'm trying to store a couple of matrices of
1 has type Int. If you add it to something with a different type, you might be
causing type instability. What happens if you replace the literal 1 with one(T)
for the type you're working with?
-- John
On Sep 18, 2014, at 9:56 AM, G. Patrick Mauroy gpmau...@gmail.com wrote:
Profiling shows
I think that was a typo for not surprised.
-- John
On Sep 18, 2014, at 9:59 AM, Steven G. Johnson stevenj@gmail.com wrote:
On Thursday, September 18, 2014 12:00:32 PM UTC-4, Florian Oswald wrote:
well, I guess most computer scientists would be surprised. writing on a piece
of paper
Try findnz.
This seems to not be documented in the sparse section of the manual, but I
would think it should be.
— John
On Sep 18, 2014, at 6:58 PM, DumpsterDoofus peter.richter@gmail.com wrote:
Given column vectors I, J, and V, one can construct a sparse matrix using the
following
401 - 500 of 726 matches
Mail list logo