allow hashing DataArrays with NA values

JuliaStats · Jun 15, 2015 · d4fdfbc · d4fdfbc
1 parent b5d39c3
commit d4fdfbc
Show file tree

Hide file tree

Showing 2 changed files with 11 additions and 4 deletions.
diff --git a/src/dataarray.jl b/src/dataarray.jl
@@ -634,10 +634,14 @@ end
 #' k = hash(dv)
 #
 # TODO: Make sure this agrees with is_equals()
-function Base.hash(a::AbstractDataArray) # -> UInt
-    h = hash(size(a)) + 1
-    for i in 1:length(a)
-        h = hash(@compat(Int(hash(a[i]))), h)
+function Base.hash(a::DataArray) # -> UInt
+    # hash NA pattern
+    h = hash(a.na)
+    # hash non-NA elements
+    i = findfirst(a.na, false)
+    while i > 0
+      h = hash(a.data[i], h)
+      i = findnext(a.na, false, i+1)
     end
     return @compat UInt(h)
 end

diff --git a/test/newtests/dataarray.jl b/test/newtests/dataarray.jl
@@ -330,4 +330,7 @@ module TestDataArrays
     hash(DataArray([1, 2], falses(2)))
     hash(DataArray(repeat([1, 2], outer = [1, 2]), falses(2, 2)))
     hash(DataArray(repeat([1, 2], outer = [1, 2, 2]), falses(2, 2, 2)))
+    hash(@data [1, NA])
+    hash(@data repeat( [1, 2, NA], outer = [1, 2]))
+    hash(@data repeat( [NA, NA, NA], outer = [1, 2, 2]), falses(2, 2, 2))
 end