diff --git a/onadata/libs/tests/utils/test_csv_import.py b/onadata/libs/tests/utils/test_csv_import.py
index 0ca4495da4..70f12a6615 100644
--- a/onadata/libs/tests/utils/test_csv_import.py
+++ b/onadata/libs/tests/utils/test_csv_import.py
@@ -1,5 +1,6 @@
 from __future__ import unicode_literals
 
+import json
 import os
 import re
 from builtins import open
@@ -21,6 +22,7 @@
 from onadata.libs.utils.common_tags import IMPORTED_VIA_CSV_BY
 from onadata.libs.utils.csv_import import get_submission_meta_dict
 from onadata.libs.utils.user_auth import get_user_default_project
+from onadata.libs.utils.csv_import import get_columns_by_type
 
 
 def strip_xml_uuid(s):
@@ -523,3 +525,31 @@ def test_csv_import_with_overwrite(self):
 
         self.assertEqual(count, 1)
         self.assertEqual(self.xform.num_of_submissions, 1)
+
+    def test_get_columns_by_type(self):
+        """
+        Test get_columns_by_type() returns columns in groups
+        """
+        self.xls_file_path = os.path.join(
+            self.fixtures_dir, "form_with_multiple_select.xlsx"
+        )
+        self._publish_xls_file(self.xls_file_path)
+        xform = XForm.objects.get()
+        columns = get_columns_by_type(["date"], json.loads(xform.json))
+        self.assertEqual(
+            columns, ["section_A/date_of_survey", "section_B/year_established"]
+        )
+        good_csv = open(
+            os.path.join(
+                self.fixtures_dir, "csv_import_with_multiple_select.csv"
+            ), "rb"
+        )
+        csv_import.submit_csv(self.user.username, xform, good_csv)
+        self.assertEqual(Instance.objects.count(), 1)
+        submission = Instance.objects.first()
+        self.assertEqual(submission.status, "imported_via_csv")
+        self.assertEqual(submission.json["section_A/date_of_survey"],
+                         "2015-09-10")
+        self.assertTrue(
+            submission.json["section_B/year_established"].startswith("1890")
+        )
diff --git a/onadata/libs/utils/csv_import.py b/onadata/libs/utils/csv_import.py
index 211c83d8d0..f25a7cec52 100644
--- a/onadata/libs/utils/csv_import.py
+++ b/onadata/libs/utils/csv_import.py
@@ -611,10 +611,26 @@ def get_columns_by_type(type_list, form_json):
     within the type_list
     :rtype: list
     """
-    return [
-        dt.get('name') for dt in form_json.get('children')
-        if dt.get('type') in type_list
-    ]
+
+    def _column_by_type(item_list, prefix=""):
+        found = []
+        for item in item_list:
+            if item["type"] in ["group", "repeat"]:
+                prefix = "/".join(
+                    [prefix, item["name"]]
+                ) if prefix else item["name"]
+                found.extend(_column_by_type(item["children"], prefix))
+                prefix = ""  # Reset prefix to blank
+            else:
+                if item["type"] in type_list:
+                    name = "%s/%s" % (
+                        prefix, item["name"]
+                    ) if prefix else item["name"]
+                    found.append(name)
+
+        return found
+
+    return _column_by_type(form_json["children"])
 
 
 def validate_row(row, columns):