diff --git a/src/calibre/ebooks/metadata/book/base.py b/src/calibre/ebooks/metadata/book/base.py
index 77df6b00c2..17f2c6705c 100644
--- a/src/calibre/ebooks/metadata/book/base.py
+++ b/src/calibre/ebooks/metadata/book/base.py
@@ -16,6 +16,7 @@ from calibre.ebooks.metadata.book import TOP_LEVEL_CLASSIFIERS
 from calibre.ebooks.metadata.book import ALL_METADATA_FIELDS
 from calibre.library.field_metadata import FieldMetadata
 from calibre.utils.date import isoformat, format_date
+from calibre.utils.icu import sort_key
 from calibre.utils.formatter import TemplateFormatter
 
 
@@ -490,7 +491,7 @@ class Metadata(object):
         return authors_to_string(self.authors)
 
     def format_tags(self):
-        return u', '.join([unicode(t) for t in self.tags])
+        return u', '.join([unicode(t) for t in sorted(self.tags, key=sort_key)])
 
     def format_rating(self):
         return unicode(self.rating)
@@ -530,7 +531,7 @@ class Metadata(object):
             orig_res = res
             datatype = cmeta['datatype']
             if datatype == 'text' and cmeta['is_multiple']:
-                res = u', '.join(res)
+                res = u', '.join(sorted(res, key=sort_key))
             elif datatype == 'series' and series_with_index:
                 if self.get_extra(key) is not None:
                     res = res + \
@@ -560,7 +561,7 @@ class Metadata(object):
             elif key == 'series_index':
                 res = self.format_series_index(res)
             elif datatype == 'text' and fmeta['is_multiple']:
-                res = u', '.join(res)
+                res = u', '.join(sorted(res, key=sort_key))
             elif datatype == 'series' and series_with_index:
                 res = res + ' [%s]'%self.format_series_index()
             elif datatype == 'datetime':
diff --git a/src/calibre/library/caches.py b/src/calibre/library/caches.py
index 2596b494bf..980c9f1fa9 100644
--- a/src/calibre/library/caches.py
+++ b/src/calibre/library/caches.py
@@ -691,13 +691,7 @@ class ResultCache(SearchQueryParser): # {{{
             fields = [('timestamp', False)]
 
         keyg = SortKeyGenerator(fields, self.field_metadata, self._data)
-        # For efficiency, the key generator returns a plain value if only one
-        # field is in the sort field list. Because the normal cmp function will
-        # always assume asc, we must deal with asc/desc here.
-        if len(fields) == 1:
-            self._map.sort(key=keyg, reverse=not fields[0][1])
-        else:
-            self._map.sort(key=keyg)
+        self._map.sort(key=keyg)
 
         tmap = list(itertools.repeat(False, len(self._data)))
         for x in self._map_filtered:
@@ -730,8 +724,6 @@ class SortKeyGenerator(object):
 
     def __call__(self, record):
         values = tuple(self.itervals(self.data[record]))
-        if len(values) == 1:
-            return values[0]
         return SortKey(self.orders, values)
 
     def itervals(self, record):
@@ -754,6 +746,11 @@ class SortKeyGenerator(object):
                     val = (self.string_sort_key(val), sidx)
 
             elif dt in ('text', 'comments', 'composite', 'enumeration'):
+                if val:
+                    sep = fm['is_multiple']
+                    if sep:
+                        val = sep.join(sorted(val.split(sep),
+                                              key=self.string_sort_key))
                 val = self.string_sort_key(val)
 
             elif dt == 'bool':