From b5cd84ae7cc3c5825311450ab7200d7a74dc6f73 Mon Sep 17 00:00:00 2001
From: Alexey Stukalov <astukalov@gmail.com>
Date: Sun, 21 Mar 2021 14:37:10 +0100
Subject: [PATCH] add generic vector conversion jlvec() method

so that it is called to convert dataframe column with non-scalar eltype.
Also restrict the first jlvec() method to numeric eltypes.

fixes #82
---
 src/convert.jl               |  11 ++++++++---
 test/RDA.jl                  |  15 +++++++++++++++
 test/data_v2/list_of_vec.rda | Bin 0 -> 228 bytes
 test/data_v3/list_of_vec.rda | Bin 0 -> 238 bytes
 test/generate_rda.R          |   8 ++++++++
 5 files changed, 31 insertions(+), 3 deletions(-)
 create mode 100644 test/data_v2/list_of_vec.rda
 create mode 100644 test/data_v3/list_of_vec.rda

diff --git a/src/convert.jl b/src/convert.jl
index 276ba9b..0d6167e 100644
--- a/src/convert.jl
+++ b/src/convert.jl
@@ -25,7 +25,7 @@ isna(x::ComplexF64) = isna(real(x)) || isna(imag(x))
 # convert R vector into Vector holding elements of type T
 # if force_missing is true, the result is always Vector{Union{T,Missing}},
 # otherwise it's Vector{T} if `rv` doesn't contain NAs
-function jlvec(::Type{T}, rv::RVEC, force_missing::Bool=true) where T
+function jlvec(::Type{T}, rv::RVEC, force_missing::Bool=true) where T <: Number
     anyna = any(isna, rv.data)
     if force_missing || anyna
         res = convert(Vector{Union{T,Missing}}, rv.data)
@@ -154,6 +154,11 @@ function jlvec(::Type{ZonedDateTime}, rv::RVEC, force_missing::Bool=true)
     return datetimes
 end
 
+# generic vector conversion
+function jlvec(::Type{T}, rv::RVEC, force_missing::Bool=true) where T
+    return sexp2julia.(rv.data)
+end
+
 function sexp2julia(rex::RSEXPREC)
     @warn "Conversion of $(typeof(rex)) to Julia is not implemented"
     return nothing
@@ -188,10 +193,10 @@ function sexp2julia(rl::RList)
         DataFrame(Any[isa(col, RAltRep) ? sexp2julia(col) : jlvec(col, false) for col in rl.data],
                   identifier.(names(rl)), makeunique=true)
     elseif hasnames(rl)
-        DictoVec(Any[sexp2julia(item) for item in rl.data], names(rl))
+        DictoVec(jlvec(Any, rl), names(rl))
     else
         # FIXME return DictoVec if forceDictoVec is on
-        map(sexp2julia, rl.data)
+        jlvec(Any, rl)
     end
 end
 
diff --git a/test/RDA.jl b/test/RDA.jl
index 9f63cf7..1f1fe7f 100644
--- a/test/RDA.jl
+++ b/test/RDA.jl
@@ -98,6 +98,21 @@ using RData
         @test f["x"] == f["y"] == ["a", "b", "c"]
     end
 
+    @testset "List of vectors (#82)" begin
+        f = load(joinpath(rdata_path, "list_of_vec.rda"))
+        @test f["listofvec"] isa Vector
+        @test length(f["listofvec"]) == 3
+        @test isequal(f["listofvec"], [[1., 2., missing], [3., 4.], [5., 6., missing]])
+
+        @test f["namedlistofvec"] isa DictoVec
+        @test length(f["namedlistofvec"]) == 3
+        @test f["namedlistofvec"].name2index == Dict("A"=>1, "B"=>3)
+        @test isequal(values(f["namedlistofvec"]), [[1., 2., missing], [3., 4.], [5., 6., missing]])
+
+        @test f["testdf"] isa DataFrame
+        @test nrow(f["testdf"]) == 3
+        @test isequal(f["testdf"][!, "listascol"], [[1., 2., missing], [3., 4.], [5., 6., missing, 7.]])
+    end
 end # for ver in ...
 
 @testset "Loading AltRep-containing RData files (version=3)" begin
diff --git a/test/data_v2/list_of_vec.rda b/test/data_v2/list_of_vec.rda
new file mode 100644
index 0000000000000000000000000000000000000000..de72cc20b3f6295f35309469a0844b687ab27451
GIT binary patch
literal 228
zcmV<A02}`wiwFP!000000}FDAFye~fVqjokVqgJ6CT1X$g$YP70(qQ3nlmS}xFkQV
zEHxR(5e8zA3O*2R|A7Gv93T`!J($nF2rAFyzyXyPV1NOzJ_iw~xCFXB6pQ)t5_3~i
zkZfn7r|qmD+lzq=0kEYIK}JU?g8?Ss1SI~UX=N)(EiOq(18Nfnhbx*Hi7-YIj6toC
zWx?e)2N{?$kVBdOL!HEg?j&$rCl)8?=b$HoqWp3_Scvd~v^D^V|NsC0MGv;*oWx>S
eu)zemQW8rN_0o!f`XRx^@DBiZnPDGz0ssK6<63zD

literal 0
HcmV?d00001

diff --git a/test/data_v3/list_of_vec.rda b/test/data_v3/list_of_vec.rda
new file mode 100644
index 0000000000000000000000000000000000000000..77b85fe8b3816bd00adb8a49eb44bb3bdcb18293
GIT binary patch
literal 238
zcmV<K01^KmiwFP!000000}FDAFy@NjVqjokW?%tAW>z4Rfi*P5P1k~ffrSakVFU_u
z0%^{i%;J*#w6fG>AV(OigMon$MB9I000RdI#ZV9CvoC_mGdXZT<pmgE0IbhJ1S&3p
zt`EgxzP!ZT)D$G!ndoUdE6Da@AVUBeNFYH*M<{~<Cg21l{-J4QD@iRbNl620138!p
z-Hb#SBMHW!R>-p8@|%MUOc}_b%>SWIVnTNkIIa_mlk;=X6G2gaxgIP;ctKhlfW-g*
o|No)~TXIffF)Y|%f?O$yC5d`zML_+K;9~d(0KdNAae)E=0Ii;42mk;8

literal 0
HcmV?d00001

diff --git a/test/generate_rda.R b/test/generate_rda.R
index 54b50f6..ceb25cb 100644
--- a/test/generate_rda.R
+++ b/test/generate_rda.R
@@ -124,6 +124,14 @@ altrepnames_df <- as.data.frame(altrepnames_list)
 names(altrepnames_df) <- names(altrepnames_list)
 save(altrepnames_list, altrepnames_df, file=file.path("data_v3", "altrep_names.rda"), version=3)
 
+# list of vectors variable and list of vectors column (for #82)
+listofvec <- list(c(1, 2, NA), c(3, 4), c(5, 6, NA))
+namedlistofvec <- list(A=c(1, 2, NA), c(3, 4), B=c(5, 6, NA))
+testdf <- data.frame(a = c("a", "b", "c"))
+testdf$listascol <- list(c(1, 2, NA), c(3, 4), c(5, 6, NA, 7))
+save(listofvec, namedlistofvec, testdf, file=file.path("data_v2", "list_of_vec.rda"), version=2)
+save(listofvec, namedlistofvec, testdf, file=file.path("data_v3", "list_of_vec.rda"), version=3)
+
 # generate files using each of the supported compression types
 df <- data.frame(num = c(1.1, 2.2))
 rdata_path <- "data_v3"