Fix NDArrayIter iteration bug when last_batch_handle='pad' (#16166)

* fix NDarrayIter pad bug

* retrigger CI

python/mxnet/io/io.py

@@ -748,7 +748,15 @@ def _batchify(self, data_source):
             self.cursor + self.batch_size > self.num_data:
             pad = self.batch_size - self.num_data + self.cursor
             first_data = self._getdata(data_source, start=self.cursor)
-            second_data = self._getdata(data_source, end=pad)
+            if pad > self.num_data:
+                while True:
+                    if pad <= self.num_data:
+                        break
+                    second_data = self._getdata(data_source, end=self.num_data)
+                    pad -= self.num_data
+                second_data = self._concat(second_data, self._getdata(data_source, end=pad))
+            else:
+                second_data = self._getdata(data_source, end=pad)
             return self._concat(first_data, second_data)
         # normal case
         else:

tests/python/unittest/test_io.py

@@ -198,6 +198,13 @@ def _test_shuffle(data, labels=None):
         assert np.array_equal(batch.data[0].asnumpy(), batch_list[idx_list[i]])
         i += 1
 
+# fixes the issue https://github.com/apache/incubator-mxnet/issues/15535
+def _test_corner_case():
+    data = np.arange(10)
+    data_iter = mx.io.NDArrayIter(data=data, batch_size=25, shuffle=False, last_batch_handle='pad')
+    expect = np.concatenate((np.tile(data, 2), np.arange(5)))
+    assert np.array_equal(data_iter.next().data[0].asnumpy(), expect)
+
 
 def test_NDArrayIter():
     dtype_list = ['NDArray', 'ndarray']
@@ -220,6 +227,7 @@ def test_NDArrayIter():
             _test_shuffle({'data1': data, 'data2': data})
             _test_shuffle(data, [])
             _test_shuffle(data)
+    _test_corner_case()
 
 
 def test_NDArrayIter_h5py():