Merge 65cc38d into c79c20d

src/routing.jl

@@ -86,14 +86,11 @@ end
 ### Create a Sparse Matrix for a given vector of weights ###
 
 function create_weights_matrix(m::OpenStreetMapX.MapData,weights::Vector{Float64})
-    w = Dict{Tuple{Int,Int},Float64}()
-    sizehint!(w,length(weights))
+    w = SparseArrays.spzeros(length(m.v), length(m.v))
     for (i,edge) in enumerate(m.e)
         w[m.v[edge[1]],m.v[edge[2]]] = weights[i]
     end
-    return SparseArrays.sparse(map(x->getfield.(collect(keys(w)), x),
-        fieldnames(eltype(collect(keys(w)))))..., 
-        collect(values(w)),length(m.v),length(m.v))
+    w
 end
 
 ### Get velocities matrix ###
@@ -102,11 +99,11 @@ function get_velocities(m::OpenStreetMapX.MapData,
             class_speeds::Dict{Int,Float64} = OpenStreetMapX.SPEED_ROADS_URBAN)
     @assert length(m.e) == length(m.w.nzval)
     indices = [(m.v[i],m.v[j]) for (i,j) in m.e]
-    V = Array{Float64}(undef,length(m.e))
-    for i = 1:length(indices)
-        V[i] = class_speeds[m.class[i]]/3.6
+    V = SparseArrays.spzeros(length(m.v), length(m.v))
+    for (i,ind) in enumerate(indices)
+        V[ind[1],ind[2]] = class_speeds[m.class[i]]/3.6
     end
-    return SparseArrays.sparse(map(i -> m.v[i[1]], m.e), map(i -> m.v[i[2]], m.e),V)
+    V
 end
 
 ### Extract route from Dijkstra results object ###