ARROW-43: Python: format array values to in __repr__ for interactive computing

python/arrow/array.pxd

@@ -34,6 +34,7 @@ cdef class Array:
         DataType type
 
     cdef init(self, const shared_ptr[CArray]& sp_array)
+    cdef getitem(self, int i)
 
 
 cdef class BooleanArray(Array):

python/arrow/array.pyx

@@ -46,6 +46,17 @@ cdef class Array:
         def __get__(self):
             return self.sp_array.get().null_count()
 
+    def __iter__(self):
+        for i in range(len(self)):
+            yield self.getitem(i)
+        raise StopIteration
+
+    def __repr__(self):
+        from arrow.formatting import array_format
+        type_format = object.__repr__(self)
+        values = array_format(self, window=10)
+        return '{0}\n{1}'.format(type_format, values)
+
     def __len__(self):
         return self.sp_array.get().length()
 
@@ -74,7 +85,10 @@ cdef class Array:
         while key < 0:
             key += len(self)
 
-        return scalar.box_arrow_scalar(self.type, self.sp_array, key)
+        return self.getitem(key)
+
+    cdef getitem(self, int i):
+        return scalar.box_arrow_scalar(self.type, self.sp_array, i)
 
     def slice(self, start, end):
         pass

python/arrow/scalar.pxd

@@ -55,7 +55,7 @@ cdef class ListValue(ArrayValue):
     cdef:
         CListArray* ap
 
-    cdef _getitem(self, int i)
+    cdef getitem(self, int i)
 
 
 cdef class StringValue(ArrayValue):

python/arrow/scalar.pyx

@@ -144,14 +144,19 @@ cdef class ListValue(ArrayValue):
         return self.ap.value_length(self.index)
 
     def __getitem__(self, i):
-        return self._getitem(i)
+        return self.getitem(i)
+
+    def __iter__(self):
+        for i in range(len(self)):
+            yield self.getitem(i)
+        raise StopIteration
 
     cdef void _set_array(self, const shared_ptr[CArray]& sp_array):
         self.sp_array = sp_array
         self.ap = <CListArray*> sp_array.get()
         self.value_type = box_data_type(self.ap.value_type())
 
-    cdef _getitem(self, int i):
+    cdef getitem(self, int i):
         cdef int j = self.ap.offset(self.index) + i
         return box_arrow_scalar(self.value_type, self.ap.values(), j)
 
@@ -161,7 +166,7 @@ cdef class ListValue(ArrayValue):
             list result = []
 
         for j in range(len(self)):
-            result.append(self._getitem(j).as_py())
+            result.append(self.getitem(j).as_py())
 
         return result
 

python/arrow/tests/test_array.py

@@ -17,10 +17,47 @@
 
 from arrow.compat import unittest
 import arrow
+import arrow.formatting as fmt
 
 
 class TestArrayAPI(unittest.TestCase):
 
     def test_getitem_NA(self):
         arr = arrow.from_pylist([1, None, 2])
         assert arr[1] is arrow.NA
+
+    def test_list_format(self):
+        arr = arrow.from_pylist([[1], None, [2, 3]])
+        result = fmt.array_format(arr)
+        expected = """\
+[
+  [1],
+  NA,
+  [2,
+   3]
+]"""
+        assert result == expected
+
+    def test_string_format(self):
+        arr = arrow.from_pylist(['foo', None, 'bar'])
+        result = fmt.array_format(arr)
+        expected = """\
+[
+  'foo',
+  NA,
+  'bar'
+]"""
+        assert result == expected
+
+    def test_long_array_format(self):
+        arr = arrow.from_pylist(range(100))
+        result = fmt.array_format(arr, window=2)
+        expected = """\
+[
+  0,
+  1,
+  ...
+  98,
+  99
+]"""
+        assert result == expected