v2.8 - faster serialization of dict's without use_strict_order

Dronehub · Aug 6, 2021 · 2e07151 · 2e07151
1 parent 02131b2
commit 2e07151
Show file tree

Hide file tree

Showing 5 changed files with 27 additions and 18 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
diff --git a/docs/changelog.md b/docs/changelog.md
@@ -1,6 +1,12 @@
 Changelog
 =========
 
+v2.8
+----
+
+* serializing a dict without :code:`use_strict_order` won't construct a list of it's items,
+    allowing you to serialize large dictionaries
+
 v2.7
 ----
 

diff --git a/docs/usage.rst b/docs/usage.rst
@@ -87,5 +87,6 @@ There's also a class available for encoding. Use it like you would a normal Pyth
 and sort them before dumping them to binary output. By enabling this feature you guarantee
 that identical dicts, serialized by identical Pythons will have the exact same binary representation.
 
-
+So take care when serializing that with large dicts, :code:`use_strict_order` will construct a list
+of all it's items, while serializing a normal dict won't.
 Only then strict order will be guaranteed. Your keys must be comparable anyway.
diff --git a/minijson.pyx b/minijson.pyx
@@ -514,13 +514,18 @@ cdef class MiniJSONEncoder:
                     cio.write(b'\x12')
                     cio.write(STRUCT_L.pack(length))
                     length = 5
-                items = list(data.items())
                 if self.use_strict_order:
+                    items = list(data.items())
                     items.sort()
-                for field_name, elem in items:
-                    cio.write(bytearray([len(field_name)]))
-                    cio.write(field_name.encode('utf-8'))
-                    length += self.dump(elem, cio)
+                    for field_name, elem in items:
+                        cio.write(bytearray([len(field_name)]))
+                        cio.write(field_name.encode('utf-8'))
+                        length += self.dump(elem, cio)
+                else:
+                    for field_name, elem in data.items():
+                        cio.write(bytearray([len(field_name)]))
+                        cio.write(field_name.encode('utf-8'))
+                        length += self.dump(elem, cio)
                 return length
             else:
                 if length <= 0xF:
@@ -537,13 +542,16 @@ cdef class MiniJSONEncoder:
                     cio.write(b'\x13')
                     cio.write(STRUCT_L.pack(length))
                     offset = 5
-
-                items = list(data.items())
                 if self.use_strict_order:
+                    items = list(data.items())
                     items.sort()
-                for key, value in items:
-                    offset += self.dump(key, cio)
-                    offset += self.dump(value, cio)
+                    for key, value in items:
+                        offset += self.dump(key, cio)
+                        offset += self.dump(value, cio)
+                else:
+                    for key, value in data.items():
+                        offset += self.dump(key, cio)
+                        offset += self.dump(value, cio)
                 return offset
         else:
             v = self.default(data)

diff --git a/setup.cfg b/setup.cfg
@@ -1,6 +1,6 @@
 # coding: utf-8
 [metadata]
-version = 2.7
+version = 2.8
 name = minijson
 long_description = file: README.md
 long_description_content_type = text/markdown; charset=UTF-8