Changes between Version 2 and Version 3 of ModelInheritance

Timestamp:

Jan 26, 2006, 12:23:50 PM (18 years ago)

Author:

jkocherhans

Comment:

initial draft of subtyping proposal... still *really* rough

ModelInheritance

@@ -2 +2 @@
 
 http://www.djangoproject.com/documentation/models/subclassing/
+
+
+= magic-removal: Model Inheritance =
+
+This is a proposal for how subclassing should work in Django. There a lot of details to get right, so this proposal should be very specific and detailed. Most of the ideas here come from the thread linked below:
+
+http://groups.google.com/group/django-developers/browse_frm/thread/ea5e0bf903058fac/9a68ac0d99cb6d7d?q=semantics&rnum=1#9a68ac0d99cb6d7d
+
+== Subclassing ==
+
+For subclassing, there are 3 main issues:
+
+ 1. How do we model the relations in SQL?
+ 2. How do joins work?
+ 3. How does the API work?
+
+Note that I have only provided examples for single inheritance here. Is multiple inheritance worth supporting?
+
+The examples will use the following models:
+{{{
+#!python
+class Place(models.Model):
+    name = models.CharField(maxlength=50)
+
+class Restaurant(Place):
+    description = models.TextField()
+
+class ItalianRestaurant(Restaurant):
+    has_decent_gnocci = models.BooleanField()
+}}}
+
+
+=== 1. Modeling parent relations in SQL? ===
+
+  1. Foreign key to parent table named $(parenttable)_id
+  2. Foreign key to parent table named 'parent_id'
+  3. 'id' of parent row equals 'id' of child row
+
+1. and 2. would take up part of the objects namespace. With 1, you couldn't have attributes named $(parenttable)_id or 'parent_id' respectively.
+
+With option 3, auto-incrementing might be harder, but it makes more sense to have 1 id across the board. I think one would expect {{{Place.objects.get(2)}}} to return the "same" object as {{{Restaurant.objects.get(2)}}}.
+
+Here's an example of option 1:
+{{{
+#!sql
+CREATE TABLE "myapp_place" (
+    "id" integer NOT NULL PRIMARY KEY,
+    "name" varchar(50) NOT NULL
+);
+
+CREATE TABLE "myapp_restaurant" (
+    "id" integer NOT NULL PRIMARY KEY,
+    "place_id" integer NOT NULL REFERENCES "myapp_place" ("id"),
+    "description" text NOT NULL
+);
+
+CREATE TABLE "myapp_italianrestaurant" (
+    "id" integer NOT NULL PRIMARY KEY,
+    "restaurant_id" integer NOT NULL REFERENCES "myapp_restaurant" ("id"),
+    "has_decent_gnocci" bool NOT NULL
+);
+}}}
+
+and option 3:
+{{{
+#!sql
+CREATE TABLE "myapp_place" (
+    "id" integer NOT NULL PRIMARY KEY,
+    "name" varchar(50) NOT NULL
+);
+
+CREATE TABLE "myapp_restaurant" (
+    "id" integer NOT NULL PRIMARY KEY, /* Could this reference myapp_place's id too? Doubtful. :( */
+    "description" text NOT NULL
+);
+
+CREATE TABLE "myapp_italianrestaurant" (
+    "id" integer NOT NULL PRIMARY KEY, /* Ditto for referencing myapp_restaurant's id */
+    "has_decent_gnocci" bool NOT NULL
+);
+}}}
+
+=== 2. Modeling joins in SQL ===
+
+When we want a list of {{{ItalianRestaurant}}}s, we obviously need all the fields from myapp_restaurant and myapp_place as well. This could be accomplished by left joins. It would look something like this:
+{{{
+#!sql
+SELECT ...
+FROM myapp_italianrestaurant as ir
+LEFT JOIN myapp_restaurant as r
+ON ir.restaurant_id=r.id
+LEFT JOIN myapp_place as p
+ON r.place_id=p.id
+}}}
+
+
+But what if we want a list of {{{Place}}}s, what should we do? We can either ''just'' get the places:
+{{{
+#!sql
+SELECT ...
+FROM myapp_place
+}}}
+
+Or we can get ''everything'' with left joins (this allows the iterator to return objects of the appropriate type, rather than just a bunch of 
+{{{Places}}}):
+{{{
+#!sql
+SELECT ...
+FROM myapp_place as p
+LEFT JOIN myapp_restaurant as r
+ON r.place_id=p.id
+LEFT JOIN myapp_italianrestaurant as ir
+ON ir.restaurant_id=r.id
+}}}
+
+Imagine we have more than one subclass of {{{Place}}} though. The join clause and the column list would get pretty hefty. This could obviously get unmanageable pretty quickly.
+
+Another option is to lazily load objects like {{{Restaurant}}} and {{{ItalianRestaurant}}} while we're iterating over {{{Place.objects.all()}}}, but that requires a ''lot'' of database queries. Either way, doing this will be expensive, and api should reflect that. You're much better off ''just'' using {{{Place}}}s fields if you are going to iterate over {{{Place.objects.all()}}}
+
+
+=== 3. API ===
+
+The following API examples assume we have created these objects:
+
+{{{
+#!python
+p = Place(name='My House')
+r = Restaurant(name='Road Kill Cafe', description='Yuck!')
+ir = ItalianRestaurant(name='Luigis', description='Something', has_decent_gnocci=False)
+}}}
+
+For the following examples, assume {{{Place.objects.get(2)}}} returns {{{r}}} and 
+{{{Place.objects.get(3)}}} returns {{{ir}}}.
+
+||    || python                                    || result ||
+|| A. || {{{Place.objects.count()}}}               || 3 ||
+|| B. || {{{Restaurant.objects.count()}}}          || 2   ||
+|| C. || {{{ItalianRestaurant.objects.count()}}}   || 1   ||
+|| D. || {{{Place.objects.get(2).description}}}    || 'Yuck!' or {{{AttributeError}}}? ||
+|| E. || {{{Restaurant.objects.get(2).description}}}    || 'Yuck!' ||
+
+
+
+== Change the current usage of subclassing ==
+
+{{{
+#!python
+class MyArticle(Article):
+  ...fields...
+  class META:
+        module_name = 'my_articles'
+        remove_fields = ...some fields...
+}}}
+
+would change to:
+{{{
+#!python
+
+class MyArticle(meta.Model):
+   ...fields...
+   class META:
+        copy_from = Article
+        remove_fields = ...some fields...
+}}}