Add a kwarg constructor for DataFrames (JuliaData#228)

Example: DataFrame(x = [1:4], y = rand(4))

src/dataframe.jl

@@ -42,8 +42,15 @@ end
 ##
 ##############################################################################
 
-# The empty DataFrame
-DataFrame() = DataFrame({}, Index())
+# A DataFrame from keyword arguments
+# This also covers the empty DataFrame.
+function DataFrame(;kwargs...)
+    result = DataFrame({}, Index())
+    for (k,v) in kwargs
+        result[string(k)] = v
+    end
+    return result
+end
 
 # No-op given a DataFrame
 DataFrame(df::DataFrame) = df

test/data.jl

@@ -182,12 +182,16 @@ df3 = DataFrame({dvint})
 df4 = DataFrame([1:4 1:4])
 df5 = DataFrame({DataVector[1,2,3,4], dvstr})
 df6 = DataFrame({dvint, dvint, dvstr}, ["A", "B", "C"])
+df7 = DataFrame(x = dvint, y = dvstr)
+@assert size(df7) == (4, 2)
+@assert isequal(df7["x"], dvint)
 
 test_group("description functions")
 @assert nrow(df6) == 4
 @assert ncol(df6) == 3
 @assert all(colnames(df6) .== ["A", "B", "C"])
 @assert all(colnames(df2) .== ["x1", "x2"])
+@assert all(colnames(df7) .== ["x", "y"])
 
 test_group("ref")
 @assert df6[2, 3] == "two"