Tables interface for FMUSolution

  - Column access defined
  - Restricting  JLD2 to version smaller 0.4.37

src/FMI.jl

@@ -181,16 +181,17 @@ function __init__()
         import .JLD2
         include("extensions/JLD2.jl")
     end
-    @require DataFrames="a93c6f00-e57d-5684-b7b6-d8193f3e46c0" begin
-        import .DataFrames
-        @require CSV="336ed68f-0bac-5ca0-87d4-7b16caf5d00b" begin
-            import .CSV
-            include("extensions/CSV.jl")   
-        end
+    @require CSV="336ed68f-0bac-5ca0-87d4-7b16caf5d00b" begin
+        import .CSV
+        include("extensions/CSV.jl")
     end
     @require MAT="23992714-dd62-5051-b70f-ba57cb901cac" begin
         import .MAT
-        include("extensions/MAT.jl")   
+        include("extensions/MAT.jl")
+    end
+    @require Tables="bd369af6-aec1-5ad0-b16a-f7cc5008161c" begin
+        import .Tables
+        include("extensions/Tables.jl")
     end
 end
 

src/extensions/CSV.jl

@@ -6,14 +6,14 @@
 using FMIImport: FMUSolution
 
 """
-ToDo: DocString.
+    fmiSaveSolutionCSV(solution::FMUSolution, filepath::AbstractString)
 
 Saves a FMUSolution to a csv file.
+
+# Arguments
+- `solution::FMUSolution`: The simulation results that should be saved
+- `filepath::AbstractString`: The path specifing where to save the results, also indicating the file format. Supports *.mat, *.csv, *.JLD2
 """
-function fmiSaveSolutionCSV(solution::FMUSolution, filepath::AbstractString) 
-    df = DataFrames.DataFrame(time = solution.values.t)
-    for i in 1:length(solution.values.saveval[1])
-        df[!, Symbol(fmi2ValueReferenceToString(solution.component.fmu, solution.valueReferences[i]))] = [val[i] for val in solution.values.saveval]
-    end
-    CSV.write(filepath, df)
+function fmiSaveSolutionCSV(solution::FMUSolution, filepath::AbstractString)
+    CSV.write(filepath, solution)
 end

src/extensions/JLD2.jl

@@ -21,4 +21,4 @@ Loads a FMUSolution. Returns a previously saved `FMUSolution`.
 """
 function fmiLoadSolutionJLD2(filepath::AbstractString; keyword="solution")
     return JLD2.load(filepath, keyword)
-end
+end

src/extensions/MAT.jl

@@ -20,6 +20,6 @@ function fmiSaveSolutionMAT(solution::FMUSolution, filepath::AbstractString)
         MAT.write(file, replace(fmi2ValueReferenceToString(solution.component.fmu, solution.valueReferences[i-1])[1], "." => "_"), v[:,i])
         # df[!, Symbol(fmi2ValueReferenceToString(solution.component.fmu, solution.valueReferences[i]))] = [val[i] for val in solution.values.saveval]
     end
-    
+
     MAT.close(file)
-end
+end

src/extensions/Tables.jl

@@ -0,0 +1,55 @@
+#
+# Copyright (c) 2023 Andreas Heuermann
+# Licensed under the MIT license. See LICENSE file in the project root for details.
+#
+
+using FMIImport: FMUSolution
+
+# Declare that FMUSolution is a table
+Tables.istable(::Type{<:FMUSolution}) = true
+
+# TODO: Define [optional] schema
+
+# Column interface
+Tables.columnaccess(table) = true
+
+function Tables.columns(solution::FMUSolution)
+    return solution
+end
+
+"""
+Retrieve a column by index.
+"""
+function Tables.getcolumn(solution::FMUSolution, i::Int)::Vector{Float64}
+    if i == 1 # Time
+        return solution.values.t
+    end
+    # Variables
+    return [val[i-1] for val in solution.values.saveval]
+end
+
+"""
+Retrieve a column by name.
+"""
+function Tables.getcolumn(solution::FMUSolution, nm::Symbol)
+    if nm == :time # Time
+        return solution.values.t
+    end
+    # Variables
+    vr = first(fmi2StringToValueReference(solution.component.fmu, string(nm)))
+    idx = findfirst(idx -> idx == vr, solution.valueReferences)
+    return [val[idx] for val in solution.values.saveval]
+end
+
+"""
+Return column names for a table as an indexable collection.
+"""
+function Tables.columnnames(solution::FMUSolution)
+    names = Symbol[]
+    push!(names, Symbol("time"))
+    for i in 1:length(solution.values.saveval[1])
+        var = fmi2ValueReferenceToString(solution.component.fmu, solution.valueReferences[i])
+        append!(names, Symbol.(var))
+    end
+    return unique(names)
+end

test/Project.toml

@@ -10,11 +10,14 @@ Logging = "56ddb016-857b-54e1-b83d-db4d58db5568"
 MAT = "23992714-dd62-5051-b70f-ba57cb901cac"
 Plots = "91a5bcdd-55d7-5caf-9e0b-520d859cae80"
 Random = "9a3f8284-a2c9-5f02-9a11-845980a1fd5c"
+Tables = "bd369af6-aec1-5ad0-b16a-f7cc5008161c"
 Test = "8dfed614-e22c-5e08-85e1-65c5234f0b40"
 Zygote = "e88e6eb3-aa80-5325-afca-941959d7151f"
 
 [compat]
 julia = "1.6"
+Tables = "1"
+JLD2 = "< 0.4.37, > 0.4.38"
 
 [extras]
 CPUSummary = "2a0fbf3d-bb9c-48f3-b0a9-814d99fd7ab9"