acxz okay guys I got something weird here
julia> a = 1
1
julia> A = [1]
1-element Array{Int64,1}:
1
julia> A2 = [[1, 1] [1, 1]]
2×2 Array{Int64,2}:
1 1
1 1
julia> one(a)
1
julia> one(A)
ERROR: MethodError: no method matching one(::Array{Int64,1})
Closest candidates are:
one(::Type{Missing}) at missing.jl:103
one(::BitArray{2}) at bitarray.jl:426
one(::Missing) at missing.jl:100
...
Stacktrace:
[1] top-level scope at REPL[59]:1
[2] run_repl(::REPL.AbstractREPL, ::Any) at /build/julia/src/julia-1.5.0/usr/share/julia/stdlib/v1.5/REPL/src/REPL.jl:288
julia> one(A2)
2×2 Array{Int64,2}:
1 0
0 1Ultimately I want to perform the following operation: inv(A), but since inv(A) calls one(A) which fails, inv(A) also fails.
acxz Looking at the source code of one it seems that it further calls convert(T, 1) but I can't seem to run that code with a dimensionful type. I believe I misunderstanding types here.
Daniel Lakeland > <@1awtumleef:matrix.org> okay guys I got something weird here
julia> a = 1 1 julia> A = [1] 1-element Array{Int64,1}: 1 julia> A2 = [[1, 1] [1, 1]] 2×2 Array{Int64,2}: 1 1 1 1 julia> one(a) 1 julia> one(A) ERROR: MethodError: no method matching one(::Array{Int64,1}) Closest candidates are: one(::Type{Missing}) at missing.jl:103 one(::BitArray{2}) at bitarray.jl:426 one(::Missing) at missing.jl:100 ... Stacktrace: [1] top-level scope at REPL[59]:1 [2] run_repl(::REPL.AbstractREPL, ::Any) at /build/julia/src/julia-1.5.0/usr/share/julia/stdlib/v1.5/REPL/src/REPL.jl:288 julia> one(A2) 2×2 Array{Int64,2}: 1 0 0 1Ultimately I want to perform the following operation:
inv(A), but sinceinv(A)callsone(A)which fails,inv(A)also fails.
Strange, inv works for me on an invertible matrix made like that:
julia> inv([[1, 1] [ 2, 1]])
2×2 Array{Float64,2}:
-1.0 2.0
1.0 -1.0
acxz That works for me as well, but inv([2]) doesn't work
Daniel Lakeland I assume you were expecting [0.5] ?
Daniel Lakeland right, because there's no multiplication defined for vectors
Daniel Lakeland you're thinking of [2] as a 1x1 matrix but it's not, it's a 1 vector
Daniel Lakeland I'm not sure if you can make a 1x1 matrix, and/or how
Daniel Lakeland reshape([2.0],(1,1))
Daniel Lakeland julia> inv(reshape([2.0],(1,1)))
1×1 Array{Float64,2}:
0.5
acxz I see, so maybe if I explain my usecase it will help. Essentially I need to take an inverse of a variable NxN matrix A based on user input. And the user can specify an positive natural number, including 1. If I try to use the reshape(... solution. I would have to write it as inv(reshape(A,(N,N)) which is an unnecessary operation for all N but 1.
acxz Ah wait maybe [[2]] works?
acxz julia> inv([[2]])
ERROR: MethodError: no method matching inv(::Array{Array{Int64,1},1})ah nope ☹️
Daniel Lakeland julia> @benchmark reshape(foo,(10,10))
BenchmarkTools.Trial:
memory estimate: 96 bytes
allocs estimate: 2
minimum time: 53.212 ns (0.00% GC)
median time: 55.661 ns (0.00% GC)
mean time: 69.874 ns (15.51% GC)
maximum time: 8.241 μs (98.01% GC)
samples: 10000
evals/sample: 986
Daniel Lakeland weird, I have no idea why that's all bold etc... but basically reshape is cheap, just do that
acxz whoa yeah that is cheap, thx for your help!
Daniel Lakeland patreon ... grr autocorrect
Daniel Lakeland I don't think so, it was just copy pasted from the REPL
&Adam Fu
&Adam * Funky
&Adam Are you using element or a different client?
&Adam (apologies for accidentally pressing enter a
&Adam nd sending "fu" if your client doesn't support edits)
&Adam For future reference, element uses commonmark for markup, so you can enclose content in a codeblock by putting triple-backticks around it.
Daniel Lakeland * ```
julia> @benchmark reshape(foo,(10,10))
BenchmarkTools.Trial:
memory estimate: 96 bytes
allocs estimate: 2
minimum time: 53.212 ns (0.00% GC)
median time: 55.661 ns (0.00% GC)
mean time: 69.874 ns (15.51% GC)
maximum time: 8.241 μs (98.01% GC)
samples: 10000
evals/sample: 986
```
Daniel Lakeland thanks!