diff --git a/addons/base_import/models.py b/addons/base_import/models.py
index 754467ed1c2..07a3439419d 100644
--- a/addons/base_import/models.py
+++ b/addons/base_import/models.py
@@ -24,6 +24,7 @@ class ir_import(orm.TransientModel):
         'res_model': fields.char('Model', size=64),
         'file': fields.binary('File'),
         'file_name': fields.char('File Name', size=None),
+        'file_mime': fields.char('File Type', size=None),
     }
 
     def get_fields(self, cr, uid, model, context=None,
@@ -174,19 +175,20 @@ class ir_import(orm.TransientModel):
 
         Will consume the first line of the ``rows`` iterator.
 
-        Returns either None (no title) or a dict mapping cell indices
+        Returns a pair of (None, None) if headers were not requested
+        or the list of headers and a dict mapping cell indices
         to key paths in the ``fields`` tree
 
         :param Iterator rows:
         :param dict fields:
         :param dict options:
-        :rtype: None | dict(int: list(str))
+        :rtype: (None, None) | (list(str), dict(int: list(str)))
         """
         if not options.get('headers'):
-            return None
+            return None, None
 
         headers = next(rows)
-        return dict(
+        return headers, dict(
             (index, [field['name'] for field in self._match_header(header, fields, options)] or None)
             for index, header in enumerate(headers)
         )
@@ -196,13 +198,16 @@ class ir_import(orm.TransientModel):
         fields-matching between the import's file data and the model's
         columns.
 
+        If the headers are not requested (not options.headers),
+        ``matches`` and ``headers`` are both ``False``.
+
         :param id: identifier of the import
         :param int count: number of preview lines to generate
         :param options: format-specific options.
                         CSV: {encoding, quote, separator, headers}
         :type options: {str, str, str, bool}
-        :returns: {fields, matches, preview} | {error, preview}
-        :rtype: {dict(str: dict(...)), dict(int, list(str)), list(list(str))} | {str, str}
+        :returns: {fields, matches, headers, preview} | {error, preview}
+        :rtype: {dict(str: dict(...)), dict(int, list(str)), list(str), list(list(str))} | {str, str}
         """
         (record,) = self.browse(cr, uid, [id], context=context)
         fields = self.get_fields(cr, uid, record.res_model, context=context)
@@ -210,13 +215,14 @@ class ir_import(orm.TransientModel):
         try:
             rows = self._read_csv(record, options)
 
-            match = self._match_headers(rows, fields, options)
+            headers, matches = self._match_headers(rows, fields, options)
             # Match should have consumed the first row (iif headers), get
             # the ``count`` next rows for preview
             preview = itertools.islice(rows, count)
             return {
                 'fields': fields,
-                'matches': match,
+                'matches': matches or False,
+                'headers': headers or False,
                 'preview': list(preview),
             }
         except (TypeError, UnicodeDecodeError), e:
diff --git a/addons/base_import/tests/test_cases.py b/addons/base_import/tests/test_cases.py
index 7cd7cbc5d00..00230f9a3c0 100644
--- a/addons/base_import/tests/test_cases.py
+++ b/addons/base_import/tests/test_cases.py
@@ -127,9 +127,11 @@ class test_match_headers_single(TransactionCase):
 
 class test_match_headers_multiple(TransactionCase):
     def test_noheaders(self):
-        self.assertIsNone(
+        self.assertEqual(
             self.registry('base_import.import')._match_headers(
-                [], [], {}))
+                [], [], {}),
+            (None, None)
+        )
     def test_nomatch(self):
         self.assertEqual(
             self.registry('base_import.import')._match_headers(
@@ -139,7 +141,11 @@ class test_match_headers_multiple(TransactionCase):
                 ]),
                 [],
                 {'headers': True}),
-            dict((index, None) for index in range(4)))
+            (
+                ['foo', 'bar', 'baz', 'qux'],
+                dict.fromkeys(range(4))
+            )
+        )
 
     def test_mixed(self):
         self.assertEqual(
@@ -153,12 +159,13 @@ class test_match_headers_multiple(TransactionCase):
                      ]}
                 ],
                 {'headers': True}),
-            {
+            (['foo', 'bar', 'baz', 'qux/corge'], {
                 0: None,
                 1: ['bar'],
                 2: ['bob'],
                 3: ['qux', 'corge'],
             })
+        )
 
 import base64, csv
 class test_preview(TransactionCase):
@@ -216,27 +223,22 @@ class test_preview(TransactionCase):
             'headers': True,
         })
 
-        p = {
-            'matches': {0: ['name'], 1: ['somevalue'], 2: None},
-            'fields': [
-                {'id': 'id', 'name': 'id', 'string': 'External ID', 'required':False, 'fields': []},
-                {'id': 'name', 'name': 'name', 'string': 'Name', 'required':False, 'fields': []},
-                {'id': 'somevalue', 'name': 'somevalue', 'string': 'Some Value', 'required':True, 'fields': []},
-                {'id': 'othervalue', 'name': 'othervalue', 'string': 'Other Variable', 'required':False, 'fields': []},
-            ],
-            'preview': [
-                ['foo', '1', '2'],
-                ['bar', '3', '4'],
-                ['qux', '5', '6'],
-            ]
-        }
-
-        self.assertEqual(result['matches'], p['matches'])
+        self.assertEqual(result['matches'], {0: ['name'], 1: ['somevalue'], 2: None})
+        self.assertEqual(result['headers'], ['name', 'Some Value', 'Counter'])
         # Order depends on iteration order of fields_get
-        self.assertItemsEqual(result['fields'], p['fields'])
-        self.assertEqual(result['preview'], p['preview'])
+        self.assertItemsEqual(result['fields'], [
+            {'id': 'id', 'name': 'id', 'string': 'External ID', 'required':False, 'fields': []},
+            {'id': 'name', 'name': 'name', 'string': 'Name', 'required':False, 'fields': []},
+            {'id': 'somevalue', 'name': 'somevalue', 'string': 'Some Value', 'required':True, 'fields': []},
+            {'id': 'othervalue', 'name': 'othervalue', 'string': 'Other Variable', 'required':False, 'fields': []},
+        ])
+        self.assertEqual(result['preview'], [
+            ['foo', '1', '2'],
+            ['bar', '3', '4'],
+            ['qux', '5', '6'],
+        ])
         # Ensure we only have the response fields we expect
-        self.assertItemsEqual(result.keys(), ['matches', 'fields', 'preview'])
+        self.assertItemsEqual(result.keys(), ['matches', 'headers', 'fields', 'preview'])
 
 class test_convert_import_data(TransactionCase):
     """ Tests conversion of base_import.import input into data which