add support for ordering by expressions

WaVEV · timgraham · commit a84bf11d2345 · 2024-08-08T15:05:57.000-04:00
diff --git a/README.md b/README.md
@@ -123,7 +123,6 @@ Migrations for 'admin':
   supported.
 
 * Ordering a `QuerySet` by `nulls_first` or `nulls_last` isn't supported.
-  Neither is randomized ordering.
 
 - `DateTimeField` doesn't support microsecond precision, and correspondingly,
   `DurationField` stores milliseconds rather than microseconds.
diff --git a/django_mongodb/compiler.py b/django_mongodb/compiler.py
@@ -1,16 +1,18 @@
 import itertools
 from collections import defaultdict
 
+from bson import SON
 from django.core.exceptions import EmptyResultSet, FullResultSet
 from django.db import DatabaseError, IntegrityError, NotSupportedError
 from django.db.models import Count, Expression
 from django.db.models.aggregates import Aggregate, Variance
-from django.db.models.expressions import Col, OrderBy, Value
+from django.db.models.expressions import Col, Ref, Value
 from django.db.models.functions.comparison import Coalesce
 from django.db.models.functions.math import Power
 from django.db.models.sql import compiler
-from django.db.models.sql.constants import GET_ITERATOR_CHUNK_SIZE, MULTI, ORDER_DIR, SINGLE
+from django.db.models.sql.constants import GET_ITERATOR_CHUNK_SIZE, MULTI, SINGLE
 from django.utils.functional import cached_property
+from pymongo import ASCENDING, DESCENDING
 
 from .base import Cursor
 from .query import MongoQuery, wrap_database_errors
@@ -25,6 +27,7 @@ class SQLCompiler(compiler.SQLCompiler):
     def __init__(self, *args, **kwargs):
         super().__init__(*args, **kwargs)
         self.aggregation_pipeline = None
+        self._order_by = None
 
     def _get_group_alias_column(self, expr, annotation_group_idx):
         """Generate a dummy field for use in the ids fields in $group."""
@@ -98,7 +101,7 @@ def _prepare_expressions_for_pipeline(self, expression, target, annotation_group
             replacements[sub_expr] = replacing_expr
         return replacements, group
 
-    def _prepare_annotations_for_aggregation_pipeline(self):
+    def _prepare_annotations_for_aggregation_pipeline(self, order_by):
         """Prepare annotations for the aggregation pipeline."""
         replacements = {}
         group = {}
@@ -110,6 +113,13 @@ def _prepare_annotations_for_aggregation_pipeline(self):
                 )
                 replacements.update(new_replacements)
                 group.update(expr_group)
+        for expr, _ in order_by:
+            if expr.contains_aggregate:
+                new_replacements, expr_group = self._prepare_expressions_for_pipeline(
+                    expr, None, annotation_group_idx
+                )
+                replacements.update(new_replacements)
+                group.update(expr_group)
         having_replacements, having_group = self._prepare_expressions_for_pipeline(
             self.having, None, annotation_group_idx
         )
@@ -121,9 +131,10 @@ def _get_group_id_expressions(self, order_by):
         """Generate group ID expressions for the aggregation pipeline."""
         group_expressions = set()
         replacements = {}
-        for expr, (_, _, is_ref) in order_by:
-            if not is_ref:
-                group_expressions |= set(expr.get_group_by_cols())
+        if not self._meta_ordering:
+            for expr, (_, _, is_ref) in order_by:
+                if not is_ref:
+                    group_expressions |= set(expr.get_group_by_cols())
         for expr, *_ in self.select:
             group_expressions |= set(expr.get_group_by_cols())
         having_group_by = self.having.get_group_by_cols() if self.having else ()
@@ -187,7 +198,7 @@ def _build_aggregation_pipeline(self, ids, group):
 
     def pre_sql_setup(self, with_col_aliases=False):
         extra_select, order_by, group_by = super().pre_sql_setup(with_col_aliases=with_col_aliases)
-        group, all_replacements = self._prepare_annotations_for_aggregation_pipeline()
+        group, all_replacements = self._prepare_annotations_for_aggregation_pipeline(order_by)
         # query.group_by is either:
         # - None: no GROUP BY
         # - True: group by select fields
@@ -207,6 +218,7 @@ def pre_sql_setup(self, with_col_aliases=False):
                     }
                 )
             self.aggregation_pipeline = pipeline
+        self._order_by = [expr.replace_expressions(all_replacements) for expr, _ in order_by]
         self.annotations = {
             target: expr.replace_expressions(all_replacements)
             for target, expr in self.query.annotation_select.items()
@@ -333,8 +345,13 @@ def build_query(self, columns=None):
         self.check_query()
         query = self.query_class(self)
         query.lookup_pipeline = self.get_lookup_pipeline()
-        query.order_by(self._get_ordering())
-        query.project_fields = self.get_project_fields(columns, ordering=query.ordering)
+        ordering_fields, order, need_extra_fields = self.preprocess_orderby()
+        query.project_fields = self.get_project_fields(columns, ordering_fields)
+        query.ordering = order
+        if need_extra_fields and columns is None:
+            query.extra_fields = self.get_project_fields(
+                ((name, field) for name, field in ordering_fields if name.startswith("__order"))
+            )
         where = self.get_where()
         try:
             expr = where.as_mql(self, self.connection) if where else {}
@@ -380,52 +397,6 @@ def project_field(column):
             + tuple(map(project_field, related_columns))
         )
 
-    def _get_ordering(self):
-        """
-        Return a list of (field, ascending) tuples that the query results
-        should be ordered by. If there is no field ordering defined, return
-        the standard_ordering (a boolean, needed for MongoDB "$natural"
-        ordering).
-        """
-        opts = self.query.get_meta()
-        ordering = (
-            self.query.order_by or opts.ordering
-            if self.query.default_ordering
-            else self.query.order_by
-        )
-        if not ordering:
-            return self.query.standard_ordering
-        default_order, _ = ORDER_DIR["ASC" if self.query.standard_ordering else "DESC"]
-        column_ordering = []
-        columns_seen = set()
-        for order in ordering:
-            if order == "?":
-                raise NotSupportedError("Randomized ordering isn't supported by MongoDB.")
-            if hasattr(order, "resolve_expression"):
-                # order is an expression like OrderBy, F, or database function.
-                orderby = order if isinstance(order, OrderBy) else order.asc()
-                orderby = orderby.resolve_expression(self.query, allow_joins=True, reuse=None)
-                ascending = not orderby.descending
-                # If the query is reversed, ascending and descending are inverted.
-                if not self.query.standard_ordering:
-                    ascending = not ascending
-            else:
-                # order is a string like "field" or "field__other_field".
-                orderby, _ = self.find_ordering_name(
-                    order, self.query.get_meta(), default_order=default_order
-                )[0]
-                ascending = not orderby.descending
-            column = orderby.expression.as_mql(self, self.connection)
-            if isinstance(column, dict):
-                raise NotSupportedError("order_by() expression not supported.")
-            # $sort references must not include the dollar sign.
-            column = column.removeprefix("$")
-            # Don't add the same column twice.
-            if column not in columns_seen:
-                columns_seen.add(column)
-                column_ordering.append((column, ascending))
-        return column_ordering
-
     @cached_property
     def collection_name(self):
         return self.query.get_meta().db_table
@@ -473,12 +444,29 @@ def get_project_fields(self, columns=None, ordering=None):
                 if self.query.alias_refcount[alias] and self.collection_name != alias:
                     fields[alias] = 1
             # Add order_by() fields.
-            for column, _ in ordering or []:
-                foreign_table = column.split(".", 1)[0] if "." in column else None
-                if column not in fields and foreign_table not in fields:
-                    fields[column] = 1
+            for alias, expression in ordering or []:
+                nested_entity = alias.split(".", 1)[0] if "." in alias else None
+                if alias not in fields and nested_entity not in fields:
+                    fields[alias] = expression.as_mql(self, self.connection)
         return fields
 
+    def preprocess_orderby(self):
+        fields = {}
+        result = SON()
+        need_extra_fields = False
+        idx = itertools.count(start=1)
+        for order in self._order_by or []:
+            if isinstance(order.expression, Ref):
+                fieldname = order.expression.refs
+            elif isinstance(order.expression, Col):
+                fieldname = order.expression.as_mql(self, self.connection).removeprefix("$")
+            else:
+                fieldname = f"__order{next(idx)}"
+                need_extra_fields = True
+            fields[fieldname] = order.expression
+            result[fieldname] = DESCENDING if order.descending else ASCENDING
+        return tuple(fields.items()), result, need_extra_fields
+
     def get_where(self):
         return self.where
 
diff --git a/django_mongodb/features.py b/django_mongodb/features.py
@@ -32,36 +32,9 @@ class DatabaseFeatures(BaseDatabaseFeatures):
         "lookup.tests.LookupTests.test_exact_none_transform",
         # "Save with update_fields did not affect any rows."
         "basic.tests.SelectOnSaveTests.test_select_on_save_lying_update",
-        # Order by constant not supported:
-        # AttributeError: 'Field' object has no attribute 'model'
-        "aggregation.tests.AggregateTestCase.test_annotate_values_list",
-        "aggregation.tests.AggregateTestCase.test_grouped_annotation_in_group_by",
-        "aggregation.tests.AggregateTestCase.test_non_grouped_annotation_not_in_group_by",
-        "aggregation.tests.AggregateTestCase.test_values_annotation_with_expression",
-        "annotations.tests.NonAggregateAnnotationTestCase.test_order_by_aggregate",
-        "model_fields.test_jsonfield.TestQuerying.test_ordering_grouping_by_count",
-        "ordering.tests.OrderingTests.test_default_ordering_does_not_affect_group_by",
-        "ordering.tests.OrderingTests.test_order_by_constant_value",
-        "expressions.tests.NegatedExpressionTests.test_filter",
-        "expressions_case.tests.CaseExpressionTests.test_order_by_conditional_implicit",
         # BaseExpression.convert_value() crashes with Decimal128.
         "aggregation.tests.AggregateTestCase.test_combine_different_types",
         "annotations.tests.NonAggregateAnnotationTestCase.test_combined_f_expression_annotation_with_aggregation",
-        # NotSupportedError: order_by() expression not supported.
-        "aggregation.tests.AggregateTestCase.test_aggregation_order_by_not_selected_annotation_values",
-        "db_functions.comparison.test_coalesce.CoalesceTests.test_ordering",
-        "db_functions.tests.FunctionTests.test_nested_function_ordering",
-        "db_functions.text.test_length.LengthTests.test_ordering",
-        "db_functions.text.test_strindex.StrIndexTests.test_order_by",
-        "expressions_case.tests.CaseExpressionTests.test_order_by_conditional_explicit",
-        "lookup.tests.LookupQueryingTests.test_lookup_in_order_by",
-        "ordering.tests.OrderingTests.test_order_by_expr_query_reuse",
-        "ordering.tests.OrderingTests.test_order_by_expression_ref",
-        "ordering.tests.OrderingTests.test_ordering_select_related_collision",
-        "queries.tests.Queries1Tests.test_order_by_related_field_transform",
-        "update.tests.AdvancedTests.test_update_ordered_by_inline_m2m_annotation",
-        "update.tests.AdvancedTests.test_update_ordered_by_m2m_annotation",
-        "update.tests.AdvancedTests.test_update_ordered_by_m2m_annotation_desc",
         # Pattern lookups that use regexMatch don't work on JSONField:
         # Unsupported conversion from array to string in $convert
         "model_fields.test_jsonfield.TestQuerying.test_icontains",
@@ -76,11 +49,6 @@ class DatabaseFeatures(BaseDatabaseFeatures):
         "db_functions.datetime.test_extract_trunc.DateFunctionWithTimeZoneTests.test_trunc_timezone_applied_before_truncation",
         # Length of null considered zero rather than null.
         "db_functions.text.test_length.LengthTests.test_basic",
-        # Key transforms are incorrectly treated as joins:
-        # Ordering can't span tables on MongoDB (value_custom__a).
-        "model_fields.test_jsonfield.TestQuerying.test_order_grouping_custom_decoder",
-        "model_fields.test_jsonfield.TestQuerying.test_ordering_by_transform",
-        "model_fields.test_jsonfield.TestQuerying.test_ordering_grouping_by_key_transform",
         # Wrong results in queries with multiple tables.
         "annotations.tests.NonAggregateAnnotationTestCase.test_annotation_reverse_m2m",
         "annotations.tests.NonAggregateAnnotationTestCase.test_annotation_with_m2m",
@@ -89,17 +57,9 @@ class DatabaseFeatures(BaseDatabaseFeatures):
         "expressions.test_queryset_values.ValuesExpressionsTests.test_values_list_expression",
         "expressions.test_queryset_values.ValuesExpressionsTests.test_values_list_expression_flat",
         "expressions.tests.IterableLookupInnerExpressionsTests.test_expressions_in_lookups_join_choice",
-        "ordering.tests.OrderingTests.test_order_by_grandparent_fk_with_expression_in_default_ordering",
-        "ordering.tests.OrderingTests.test_order_by_parent_fk_with_expression_in_default_ordering",
-        "ordering.tests.OrderingTests.test_order_by_ptr_field_with_default_ordering_by_expression",
         "queries.tests.Queries1Tests.test_order_by_tables",
         "queries.tests.TestTicket24605.test_ticket_24605",
         "queries.tests.TestInvalidValuesRelation.test_invalid_values",
-        # alias().order_by() doesn't work.
-        "annotations.tests.AliasTests.test_order_by_alias",
-        "annotations.tests.AliasTests.test_order_by_alias_aggregate",
-        # annotate() + values_list() + order_by() loses annotated value.
-        "expressions_case.tests.CaseExpressionTests.test_annotate_values_not_in_order_by",
         # QuerySet.explain() not implemented:
         # https://github.com/mongodb-labs/django-mongodb/issues/28
         "queries.test_explain.ExplainUnsupportedTests.test_message",
@@ -108,9 +68,6 @@ class DatabaseFeatures(BaseDatabaseFeatures):
         "aggregation.tests.AggregateTestCase.test_aggregation_default_passed_another_aggregate",
         "aggregation.tests.AggregateTestCase.test_annotation_expressions",
         "aggregation.tests.AggregateTestCase.test_reverse_fkey_annotate",
-        # Incorrect order: pipeline does not order by the correct fields.
-        "aggregation.tests.AggregateTestCase.test_annotate_ordering",
-        "aggregation.tests.AggregateTestCase.test_even_more_aggregate",
     }
     # $bitAnd, #bitOr, and $bitXor are new in MongoDB 6.3.
     _django_test_expected_failures_bitwise = {
@@ -395,6 +352,8 @@ def django_test_expected_failures(self):
         "Cannot use QuerySet.update() when querying across multiple collections on MongoDB.": {
             "queries.tests.Queries4Tests.test_ticket7095",
             "queries.tests.Queries5Tests.test_ticket9848",
+            "update.tests.AdvancedTests.test_update_ordered_by_m2m_annotation",
+            "update.tests.AdvancedTests.test_update_ordered_by_m2m_annotation_desc",
         },
         "QuerySet.dates() is not supported on MongoDB.": {
             "aggregation.tests.AggregateTestCase.test_dates_with_aggregation",
@@ -577,10 +536,6 @@ def django_test_expected_failures(self):
             "model_fields.test_jsonfield.TestQuerying.test_none_key",
             "model_fields.test_jsonfield.TestQuerying.test_none_key_exclude",
         },
-        "Randomized ordering isn't supported by MongoDB.": {
-            "aggregation.tests.AggregateTestCase.test_aggregation_random_ordering",
-            "ordering.tests.OrderingTests.test_random_ordering",
-        },
         "Queries without a collection aren't supported on MongoDB.": {
             "queries.test_q.QCheckTests",
             "queries.test_query.TestQueryNoModel",
diff --git a/django_mongodb/query.py b/django_mongodb/query.py
@@ -9,7 +9,6 @@
 from django.db.models.sql.constants import INNER
 from django.db.models.sql.datastructures import Join
 from django.db.models.sql.where import AND, OR, XOR, NothingNode, WhereNode
-from pymongo import ASCENDING, DESCENDING
 from pymongo.errors import DuplicateKeyError, PyMongoError
 
 
@@ -51,28 +50,11 @@ def __init__(self, compiler):
         self.lookup_pipeline = None
         self.project_fields = None
         self.aggregation_pipeline = compiler.aggregation_pipeline
+        self.extra_fields = None
 
     def __repr__(self):
         return f"<MongoQuery: {self.mongo_query!r} ORDER {self.ordering!r}>"
 
-    def order_by(self, ordering):
-        """
-        Reorder query results or execution order. Called by compiler during
-        query building.
-
-        `ordering` is a list with (column, ascending) tuples or a boolean --
-        use natural ordering, if any, when the argument is True and its reverse
-        otherwise.
-        """
-        if isinstance(ordering, bool):
-            # No need to add {$natural: ASCENDING} as it's the default.
-            if not ordering:
-                self.ordering.append(("$natural", DESCENDING))
-        else:
-            for column, ascending in ordering:
-                direction = ASCENDING if ascending else DESCENDING
-                self.ordering.append((column, direction))
-
     @wrap_database_errors
     def delete(self):
         """Execute a delete query."""
@@ -97,6 +79,8 @@ def get_pipeline(self):
             pipeline.extend(self.aggregation_pipeline)
         if self.project_fields:
             pipeline.append({"$project": self.project_fields})
+        if self.extra_fields:
+            pipeline.append({"$addFields": self.extra_fields})
         if self.ordering:
             pipeline.append({"$sort": dict(self.ordering)})
         if self.query.low_mark > 0:
