From 92d6d95b31da24aa58c268fc01313917e1c138a2 Mon Sep 17 00:00:00 2001
From: nyxst4ck <289980115+nyxst4ck@users.noreply.github.com>
Date: Mon, 29 Jun 2026 22:46:38 -0300
Subject: [PATCH] fix(schema): handle collection pre-load before validators

---
 AUTHORS.rst               |  1 +
 src/marshmallow/schema.py |  7 +++++--
 tests/test_decorators.py  | 16 ++++++++++++++++
 3 files changed, 22 insertions(+), 2 deletions(-)

diff --git a/AUTHORS.rst b/AUTHORS.rst
index cd928bdf9..16196f377 100644
--- a/AUTHORS.rst
+++ b/AUTHORS.rst
@@ -188,3 +188,4 @@ Contributors (chronological)
 - `@rstar327 <https://github.com/rstar327>`_
 - Kadir Can Ozden  `@bysiber  <https://github.com/bysiber>`_
 - Dhruvil Darji  `@dhruvildarji  <https://github.com/dhruvildarji>`_
+- `@nyxst4ck <https://github.com/nyxst4ck>`_
diff --git a/src/marshmallow/schema.py b/src/marshmallow/schema.py
index 8024dfe13..2c4172e7d 100644
--- a/src/marshmallow/schema.py
+++ b/src/marshmallow/schema.py
@@ -1192,9 +1192,12 @@ def _invoke_schema_validators(
             pass_original = validator_kwargs.get("pass_original", False)
 
             if many and not pass_collection:
-                for idx, (item, orig) in enumerate(
+                data_items = (
                     zip(data, original_data, strict=True)
-                ):
+                    if pass_original
+                    else ((item, None) for item in data)
+                )
+                for idx, (item, orig) in enumerate(data_items):
                     self._run_validator(
                         validator,
                         item,
diff --git a/tests/test_decorators.py b/tests/test_decorators.py
index fce7e442b..b6efd6343 100644
--- a/tests/test_decorators.py
+++ b/tests/test_decorators.py
@@ -619,6 +619,22 @@ def check(datum):
         errors = schema.validate([{"foo": 4, "bar": "42"}], many=True)
         assert errors["_schema"] == ["bar cannot be a string"]
 
+    def test_collection_pre_load_can_change_length_before_item_schema_validation(self):
+        class MySchema(Schema):
+            foo = fields.Int()
+
+            @pre_load(pass_collection=True)
+            def remove_items_without_foo(self, data, **kwargs):
+                return [item for item in data if item.get("foo") != ""]
+
+            @validates_schema
+            def validate_foo(self, data, **kwargs):
+                assert "foo" in data
+
+        schema = MySchema()
+
+        assert schema.load([{"foo": 24}, {"foo": ""}], many=True) == [{"foo": 24}]
+
     def test_allow_reporting_field_errors_in_schema_validator(self):
         class NestedSchema(Schema):
             baz = fields.Int(required=True)