add csv merge

bin/csvmerge.jl

@@ -0,0 +1,24 @@
+using ArgParse
+using UKBMain
+
+function parse_commandline()
+    s = ArgParseSettings(description="Merges 2 .csv file by SAMPLE_ID")
+
+    @add_arg_table s begin
+        "csv1"
+            help = "First .csv file"
+            arg_type = String
+        "csv2"
+            help = "Second .csv file"
+            arg_type = String
+        "out"
+            help = "Output file path"
+            arg_type = String
+    end
+
+    return parse_args(s)
+end
+
+parsed_args = parse_commandline()
+
+csvmerge(parsed_args)

src/UKBMain.jl

@@ -74,5 +74,12 @@ function decode(parsed_args)
     CSV.write(outfile, dataset)
 end
 
-export decode
+function csvmerge(parsed_args)
+    csv₁ = CSV.read(parsed_args["csv1"], DataFrame)
+    csv₂ = CSV.read(parsed_args["csv2"], DataFrame)
+    CSV.write(parsed_args["out"], innerjoin(csv₁, csv₂, on=:SAMPLE_ID))
+end
+
+export decode, csvmerge
+
 end

test/runtests.jl

@@ -82,3 +82,31 @@ end
 
     rm(parsed_args["out"])
 end
+
+@testset "Test csvmerge" begin
+    CSV. write("test_csv1.csv",
+        DataFrame(SAMPLE_ID=[1,2,3], COL1=[1., 2., 3.])
+    )
+    CSV. write("test_csv2.csv",
+        DataFrame(SAMPLE_ID=[4,3,2], COL2=[1., 2., 3.])
+    )
+
+    parsed_args = Dict(
+        "csv1" => "test_csv1.csv",
+        "csv2" => "test_csv2.csv",
+        "out" => "test_out.csv"
+    )
+    csvmerge(parsed_args)
+
+    out = CSV.read(parsed_args["out"], DataFrame)
+
+    @test out == DataFrame(
+        SAMPLE_ID = [3, 2],
+        COL1      = [3., 2.],
+        COL2      = [2., 3.]
+    )
+
+    for file in values(parsed_args)
+        rm(file)
+    end
+end