Thanks Johan, that was indeed the problem and my code runs now. :-)
I haven't dug that deep into the internals, but I tend to use the
`f1(T1::Type)` where T1 is completely unconstrained, and can be any type in
the system. On the other hand, I use the second version when I want to
constrain the types that can be possible arguments:
`f2{T2:Number}(::Type{T2})`
Short of re-implementing parseint, is there a way to work around this?
I was thinking of something like this (for Int64 only)
try
num = parseint(Int64, str)
if num != parse(str) throw(FOO) end
num
catch ... end
But I don't know if it would work properly on 32 bits CPUs.
--
Hi Henrik,
Have you checked out IJulia? https://github.com/JuliaLang/IJulia.jl
Best,
Oliver
Den fredag den 28. februar 2014 11.28.23 UTC+1 skrev Henrik L Nordmark:
Is there a way of running a Julia script line by line?
When I use R or R Studio, I am able to write an R script and run
I would love to hear an authoritative answer to this.
My belief is, that for the examples you give, the semantics should be
equivalent, but that in general f2{T2}(...) is more powerful because it
introduces a type variable (which behaves as any other variable in the
function body) that can be
Ah, great. That solves my problem.
simon
On Thursday, 27 February 2014 22:24:41 UTC, Kevin Squire wrote:
For reference, there's an open issue about this:
https://github.com/JuliaLang/julia/issues/3894
Regarding your example, you can use (or import) Main.OPTIONS to access the
value inside
On Fri, Feb 28, 2014 at 1:47 PM, Stefan Karpinski ste...@karpinski.orgwrote:
Why is the empty string the zero of the String type? The empty string is
the unit of the string monoid, so if anything, one(s::String) should give
the empty string.
Must zero(::String) != one(::String) ? :)
Not
Algebraically, the zero string should be such that concatenating
(multiplying) it with any string should yield the zero string. There is no
such string.
So we don't want to mess with the zero function for this purpose (compared
to that, even a default function would be better).
Why do you want
On Friday, 28 February 2014 10:09:20 UTC-3, Toivo Henningsson wrote:
(Related, why is :: used in parameter declarations, rather than :? I can
see that it doesn't really matter since concrete types are leaves in the
graph, and so constraints on abstract types must be : even if it says ::,
That approach works, at least for 64 bits CPUs.
In v0.2, parse(str) will choke on number overflow and throw a ParseError.
It would be nice if someone using v0.2.1 and a 32bit machine (old OS X?)
could tell me what happens when you run this:
parse(9223372036854775807)
I am trying to cut down on memory usage by using Int32 instead of the
default (on my machine) Int64 in some large data structures.
This is causing me some headaches since Int64s seem to be endemic. E.g.:
julia x = int32(1)
1
julia y = int32(2)
2
julia z = x + y
3
julia typeof(ans)
Int64
This
There was quite some discussion on this topic in
https://groups.google.com/forum/?fromgroups=#!searchin/julia-users/Int32/julia-users/Rte_I6htLRc/VJG5DWVcZbQJ
I had the feeling that we almost convinced Stefan :-)
The good thing is that Array{Int32} operations do not promote to
Array{Int64}. But
El jueves, 27 de febrero de 2014 18:45:35 UTC-6, Kevin Squire escribió:
So, no one has really explained why this is happening.
Sets are built on top of hash tables. The hash table itself is just an
array, many entries of which are not being used.
As with all iteration in Julia, for
the current behaviour is explained at
http://julia.readthedocs.org/en/latest/manual/conversion-and-promotion/
defining
Base.promote_rule(::Type{Int32}, ::Type{Int32}) = Int32
doesn't help either, and i'm not sure why.
andrew
On Friday, 28 February 2014 10:46:16 UTC-3, Tobias Knopp wrote:
El viernes, 28 de febrero de 2014 07:49:09 UTC-6, andrew cooke escribió:
the current behaviour is explained at
http://julia.readthedocs.org/en/latest/manual/conversion-and-promotion/
But I don't find it evident from that document that Int32+Int32 gives Int64
on a 64-bit machine!
On Friday, 28 February 2014 10:58:24 UTC-3, David P. Sanders wrote:
El viernes, 28 de febrero de 2014 07:49:09 UTC-6, andrew cooke escribió:
the current behaviour is explained at
I am investigating possible data structures for an application.
Here is an interesting behaviour in IntSet, which is no doubt to do with
the implementation.
Maybe it should just throw an exception if someone tries to add a really
large integer like this!
julia s = IntSet()
IntSet()
julia
The documentation states very clear that
IntSethttp://docs.julialang.org/en/latest/stdlib/base/#Base.IntSet should
only be used for dense collections, and that
Sethttp://docs.julialang.org/en/latest/stdlib/base/#Base.Set,
should be used for sparse collections.
Construct a sorted set of the
This is very nice.
Now that we have several back-ends for vectorized computation, VML,
Yeppp, Julia's builtin functions, as well as the @simd-ized versions, I am
considering whether there is a way to switch back-end without affecting the
client codes.
- Dahua
On Thursday, February 27, 2014
If you get them all to export the same API, you could, in principle, just
switch `using VML` to `using Yeppp`.
My question: are we finally conceding that add! and co. is probably worth
having?
— John
On Feb 28, 2014, at 7:10 AM, Dahua Lin linda...@gmail.com wrote:
This is very nice.
Now
I'd like to replicate the behavior of matlab's `plotyy` in Gadfly, i.e.,
I'm trying to plot data with very different `y` axes but the same `x` axes
on the same plot, any idea of how to get this done? `layer` always seems to
use the same base `y` axis.
Thanks for any pointers,
Sven
El viernes, 28 de febrero de 2014 08:41:37 UTC-6, Ivar Nesje escribió:
The documentation states very clear that
IntSethttp://docs.julialang.org/en/latest/stdlib/base/#Base.IntSet should
only be used for dense collections, and that
For the arithmetic functions, the mutating versions are probably not worth
it. For add!, subtract!, and multiply! the VML implementations on Float64
are not any faster, and with the loop vectorizer I am hoping we can erase
the remainder of the gap for Float32. In any case these operations are
Don't do it. It's not good data visualization practice, and is explicitly
and intentionally not supported in most grammar of graphics
implementations. See, for one recent post I have handy:
http://junkcharts.typepad.com/junk_charts/2014/02/a-message-worth-repeating.html
Two good alternatives are
Hi, anyone know how to help me with getting ijulia to run again (using
mavericks osx). I have clean install latest julia. I added a soft link so
captiveportal-49-129:~ jon$ which julia
/usr/local/bin/julia
and typing julia starts julia as expected from anywhere
Starting python wight he julia
While the merging of DateTime functionality into Base is simmering, I've
finally got around to making the rewrite more broadly available. You can
see the latest DateTime functionality by running:
Pkg.add(Datetime)
Pkg.checkout(Datetime,dev)
include(Pkg.dir() * /Datetime/src/Dates.jl)
using
My point was to illustrate that the limit will be arbitrary chosen. It is
impossible to set a limit for how big numbers that should be allowed to be
stored in a IntSet. It depends on the application, and can only be
determined by the programmer. Sorry for not stating that clear. You seem to
We didn't backport any of the number parsing changes to 0.2.1 since those
are features.
On Fri, Feb 28, 2014 at 8:27 AM, Pierre-Yves Gérardy pyg...@gmail.comwrote:
That approach works, at least for 64 bits CPUs.
In v0.2, parse(str) will choke on number overflow and throw a ParseError.
It
While VML is generally much faster for big arrays, the overhead is
considerable or even dominant for small ones.
I think the ideal way is to let people to explicit articulate their
intention to use these instead of changing the behavior silently. For
example, I may want to be able to use
::Type and ::Type{T} will always have the same method applicability.
The main use of the second form, as Avik says, is to constrain T.
Things get more interesting when you start nesting types. For example:
f(::Type{Array}) = 0
f{T}(::Type{Array{T}}) = 1
The rule is that the system must be able
sizeof() is a bit tricky. Currently IntSet is considered not to have a
canonical binary representation, so sizeof() defaults to giving the
size of the struct. This is a bit sketchy, but it tends to be
convenient.
On Fri, Feb 28, 2014 at 11:49 AM, Ivar Nesje iva...@gmail.com wrote:
My point was
I have benchmarked with short vectors, and for allocating functions, the
overhead of allocation dominates everything else, so it doesn't really
matter what implementation is used. I haven't tested mutating functions,
but we don't presently have those in Base. I think it would be fine to tell
It's a pity that Sublime is proprietary and not open source software.
Dear Henrik, I'm also a power Rstudio user who couldn't got used to (at the
begining) to IJulia. A few months ago I said to myself I'll wait for a
nice RStudio-like IDE and then I'll start learning it. Luckily after a few
Here is a video demo of wolfram.
http://blog.wolfram.com/2014/02/24/starting-to-demo-the-wolfram-language/
Just to be blatantly obvious about the double quotes:
julia print() # no spaces
julia print( ) # one space
julia print( ) # three spaces
Cheers!
Kevin
On Fri, Feb 28, 2014 at 10:01 AM, John Myles White johnmyleswh...@gmail.com
wrote:
This isn't a bug.
I think you want
Why this bug bother me is that, if I use print(' ') inside a function,
it won't point the error to the line where print(' ') locates but to
the line where the function starts.
It's confusing for debug!
On Friday, February 28, 2014 12:00:52 PM UTC-6, Jiakui Wang wrote:
I am on Ubuntu 13.04.
I played with this bug a little bit. If I use print(' ') inside a for
loop, the error will point to the line where for loop starts.
On Friday, February 28, 2014 12:48:04 PM UTC-6, Jacob Quinn wrote:
That's a lot of exclamation points...
Are you on windows?
Hi!
I'm doing the Machine Learning course exercises with Julia. I know how to
use Optim.jl when the cost function only have one argument, for example:
*f(t) = someFunctionOfTheta(t)*
*optimize(f, initial_theta)*
But one of the exercises is to run Octave's fminuc this way:
*options =
Yes, I tried it. My specific question is about the Base.parser() behavior
regarding integer literals:
v0.3-pre/x64 produces an Int64 when it fits, and a :(@int128_str ...)
Exprwhen confronted with numbers that would overflow an
Int64, but fit an Int128.
v0.2/x64 chokes on numbers that
There's probably a fancier way of Currying or partial application, but the
direct translation of making an anonymous function of just the first input
would be:
t - costFunction(t, X, y)
On Friday, February 28, 2014 11:18:54 AM UTC-8, Paulo Castro wrote:
Hi!
I'm doing the Machine Learning
It seems the ipython kernel manager could not find/start ijulia kernel.
Have you reinstall IJulia? Try `Pkg.add(IJulia)` in Julia REPL
On Friday, February 28, 2014 11:20:54 AM UTC-5, Jon Norberg wrote:
Hi, anyone know how to help me with getting ijulia to run again (using
mavericks osx). I
To realize what you want to do, you can add the elements to be removed to
another set.
On each iteration, verify that the current element isn't already in the set
of elements to be removed, and after the loop, clean the first set using
the elements of the second one.
—Pierre-Yves
On
El viernes, 28 de febrero de 2014 16:41:20 UTC-6, Pierre-Yves Gérardy
escribió:
To realize what you want to do, you can add the elements to be removed to
another set.
On each iteration, verify that the current element isn't already in the
set of elements to be removed, and after the
Thanks :-)
On Saturday, March 1, 2014 12:17:36 AM UTC+1, Stefan Karpinski wrote:
On Fri, Feb 28, 2014 at 2:25 PM, Pierre-Yves Gérardy
pyg...@gmail.comjavascript:
wrote:
Yes, I tried it. My specific question is about the Base.parser()behavior
regarding integer literals:
v0.3-pre/x64
Thanks Jameson. But looking in tests, I can't seem to find any examples
which constrain using of a module to a let block. Am I missing something?
Cheers,
Kevin
On Fri, Feb 28, 2014 at 3:21 PM, Jameson Nash vtjn...@gmail.com wrote:
Let blocks already do that. See the Julia tests for
Okay. I think that's a little different than what I was asking for, which
was a block structure within which I have access to a module's exports, but
external to which I don't. using ModuleX and only works on a file level,
currently.
Kevin
On Fri, Feb 28, 2014 at 4:02 PM, Jameson Nash
Hi Folks:
Newbie here...
I often need to write code that iteratively reads files which are saved from a
long MCMC chain. Coming from Matlab I always try the following
a = []
b = []
k = 1
while isfile(data$k.csv)
data = readcsv(data$k.csv)
a = [a data[:,1]]
b
I'm considering trying out a alternative IntSet implementation using a
space-optimized van Emde Boas tree. In that case, the space would scale
linearly as you add elements, and could give reasonable out-of-memory
errors should that situation arise. The limits in that case would be the
limits
Hello all, I'm trying to select a column from a DataFrame using a
variable... something like this
df = DataFrame(A=[1:10], B=[11:20])
when i do this:
df [ : 2]
i select column B
if i do this:
t = 2
df [ : t]
i get an error
Error Key not found :t
what am i missing??
Thank you in
Hi Jason,
I’m surprised that df[: 2] works: that seems almost like a bug.
To get column B, either of the following should work on a new copy of
DataFrames:
df[2]
df[:B]
If you have a slightly older copy of DataFrames (like you’d get for Julia 0.2),
this works
df[2]
df[“B”]
Hope that helps.
that does the trick, thank you very much!
On Friday, February 28, 2014 9:36:40 PM UTC-5, Jason Solack wrote:
Hello all, I'm trying to select a column from a DataFrame using a
variable... something like this
df = DataFrame(A=[1:10], B=[11:20])
when i do this:
df [ : 2]
i select column
On Saturday, March 1, 2014 12:00:29 AM UTC+11, Toivo Henningsson wrote:
Algebraically, the zero string should be such that concatenating
(multiplying) it with any string should yield the zero string. There is no
such string.
So we don't want to mess with the zero function for this purpose
Have you guys seen this? https://gitter.im
It allows setting up a persistent chat room for a git repo. Could be an
interesting alternative for IRC (logging, github-flavored markdown, nicer
design...)
It seems like a chat room for JuliaLang/julia needs to be set up by someone
with commit
53 matches
Mail list logo
