Char array fixes

src/MAT_HDF5.jl

@@ -872,14 +872,31 @@ function read(obj::Union{HDF5.Dataset,HDF5.Attribute}, ::Type{MatlabString})
         data = reshape(data, sz[2:end])
     end
     if ndims(data) == 1
-        return String(convert(Vector{Char}, data))
+        return convert_string(data)
     elseif ndims(data) == 2
-        return datap = String[rstrip(String(convert(Vector{Char}, vec(data[i, :])))) for i = 1:size(data, 1)]
+        return String[convert_string(c) for c in eachrow(data)]
     else
-        return data
+        return stringify_eachrow(data)
     end
 end
 
+function stringify_eachrow(data)
+    dims = size(data)
+    fixed_dims = [1, 3:length(dims)...] # all except 2
+    output_dims = dims[fixed_dims]
+    output = Array{String}(undef, output_dims...)
+    for I in CartesianIndices(output_dims)
+        idx = ntuple(d -> d == 2 ? Colon() : I[d < 2 ? d : d-1], length(dims))
+        slice = view(data, idx...)
+        output[I] = convert_string(slice)
+    end
+    return output
+end
+
+function convert_string(v::AbstractArray{UInt16})
+    return String(convert(Vector{Char}, v))
+end
+
 ## Utilities for handling complex numbers
 function build_datatype_complex(T::Type)
     memtype = create_datatype(HDF5.API.H5T_COMPOUND, 2*sizeof(T))

src/MAT_v5.jl

@@ -289,7 +289,7 @@ function read_string(f::IO, swap_bytes::Bool, dimensions::Vector{Int32})
         else
             data = Vector{String}(undef, dimensions[1])
             for i = 1:dimensions[1]
-                data[i] = rstrip(String(chars[i:dimensions[1]:end]))
+                data[i] = String(chars[i:dimensions[1]:end])
             end
         end
     elseif dtype <= 4 || dtype == 17
@@ -316,7 +316,7 @@ function read_string(f::IO, swap_bytes::Bool, dimensions::Vector{Int32})
         elseif dimensions[1] == 1
             data = String(take!(bufs[1]))
         else
-            data = String[rstrip(String(take!(buf))) for buf in bufs]
+            data = String[String(take!(buf)) for buf in bufs]
         end
     else
         error("Unsupported string type")
@@ -372,7 +372,7 @@ function read_matrix(f::IO, swap_bytes::Bool, subsys::Subsystem)
         data = read_struct(f, swap_bytes, dimensions, class == mxOBJECT_CLASS, subsys)
     elseif class == mxSPARSE_CLASS
         data = read_sparse(f, swap_bytes, dimensions, flags)
-    elseif class == mxCHAR_CLASS && length(dimensions) <= 2
+    elseif class == mxCHAR_CLASS
         data = read_string(f, swap_bytes, dimensions)
     elseif class == mxFUNCTION_CLASS
         data = read_matrix(f, swap_bytes, subsys)

test/read.jl

@@ -81,7 +81,7 @@ for _format in ["v6", "v7", "v7.3"]
     result = Dict(
         "simple_string" => "the quick brown fox",
         "accented_string" => "thé qüîck browñ fòx",
-        "concatenated_strings" => String["this is a string", "this is another string"],
+        "concatenated_strings" => String["this is a string      ", "this is another string"],
         "cell_strings" => Any["this is a string" "this is another string"],
         "empty_string" => ""
     )
@@ -159,6 +159,16 @@ for _format in ["v6", "v7", "v7.3"]
 
 end
 
+for _format in ["v7", "v7.3"]
+    result = Dict{String,Any}(
+        "s" => " aαβ ", # test α and β characters, not possible in v6
+        "s2" => ["fòx", "aαβ", " ef", "ac "],
+        "s3" => reshape(["faò", "aeα", " xc", "fβ "], 2, 2),
+        "s4" => Any["fòx"; "aαβ"; " ef"; "ac ";;]
+    )
+    check(joinpath(dirname(@__FILE__), _format, "char_arrays.mat"), result)
+end
+
 result = Dict(
     "index" => [8.8604784000000000e+04   9.8707212000000000e+04   1.0394035200000000e+05   1.1429712000000000e+05   1.5474923999999999e+05   1.5475435200000001e+05   1.5501074400000001e+05   1.5505315200000000e+05   1.5505718400000001e+05   1.5506168400000001e+05   1.5506524799999999e+05   5.4945741599999997e+05   5.6345896799999999e+05   5.9956981200000003e+05   7.0691623199999996e+05   7.9063657200000004e+05   8.4311938800000004e+05   9.2225131200000003e+05   1.1248994160000000e+06   1.2508148520000000e+06   1.4164141320000000e+06   1.4275988280000000e+06   1.4744331000000001e+06   1.4982212879999999e+06   1.5549058440000000e+06   1.5870300840000000e+06   1.6192005120000001e+06   1.6766071560000000e+06   1.9386816839999999e+06   1.9969427879999999e+06   2.0021861880000001e+06   2.3272494120000000e+06   2.5309351080000000e+06   2.6743788720000000e+06],
     "spikes" => [