From 2908d6dd3f68d93349b583de450bb33eaeda4dd0 Mon Sep 17 00:00:00 2001
From: Kanthi Subramanian <subkanthi@gmail.com>
Date: Thu, 12 Mar 2026 15:54:06 +0100
Subject: [PATCH] Fixed s3 decoding

Signed-off-by: Kanthi Subramanian <subkanthi@gmail.com>
---
 src/Storages/ObjectStorage/Utils.cpp          |  8 ++--
 .../integration/test_database_iceberg/test.py | 48 ++++++++++++++++++-
 2 files changed, 52 insertions(+), 4 deletions(-)

diff --git a/src/Storages/ObjectStorage/Utils.cpp b/src/Storages/ObjectStorage/Utils.cpp
index b5330ac180b2..8f490b504cc2 100644
--- a/src/Storages/ObjectStorage/Utils.cpp
+++ b/src/Storages/ObjectStorage/Utils.cpp
@@ -513,9 +513,11 @@ std::pair<DB::ObjectStoragePtr, std::string> resolveObjectStorageForPath(
             normalized_path = "gs://" + target_decomposed.authority + "/" + target_decomposed.key;
         }
         S3::URI s3_uri(normalized_path);
-
-        std::string key_to_use = s3_uri.key;
-
+      
+        // Use key (parsed without URI decoding) so that percent-encoded
+        // characters in object keys (e.g. %2F in Iceberg partition paths) are preserved.
+        std::string key_to_use = target_decomposed.key;
+      
         bool use_base_storage = false;
         if (base_storage->getType() == ObjectStorageType::S3)
         {
diff --git a/tests/integration/test_database_iceberg/test.py b/tests/integration/test_database_iceberg/test.py
index fa1ff5b45359..57efe7d2fb2e 100644
--- a/tests/integration/test_database_iceberg/test.py
+++ b/tests/integration/test_database_iceberg/test.py
@@ -625,6 +625,52 @@ def test_table_with_slash(started_cluster):
     assert node.query(f"SELECT * FROM {CATALOG_NAME}.`{root_namespace}.{table_encoded_name}`") == "\\N\tAAPL\t193.24\t193.31\t('bot')\n"
 
 
+def test_partition_value_with_slash(started_cluster):
+    """Partition value containing '/' produces object keys with %2F; reading must preserve encoding."""
+    node = started_cluster.instances["node1"]
+
+    test_ref = f"test_partition_slash_{uuid.uuid4()}"
+    table_name = f"{test_ref}_table"
+    root_namespace = f"{test_ref}_namespace"
+
+    # Partition by symbol (string) so partition value "us/west" becomes path segment symbol=us%2Fwest
+    partition_spec = PartitionSpec(
+        PartitionField(
+            source_id=2, field_id=1000, transform=IdentityTransform(), name="symbol"
+        )
+    )
+    schema = DEFAULT_SCHEMA
+
+    catalog = load_catalog_impl(started_cluster)
+    catalog.create_namespace(root_namespace)
+
+    table = create_table(
+        catalog,
+        root_namespace,
+        table_name,
+        schema,
+        partition_spec=partition_spec,
+        sort_order=DEFAULT_SORT_ORDER,
+    )
+
+    # Write a row with partition value containing slash (path will have %2F in S3 key)
+    data = [
+        {
+            "datetime": datetime.now(),
+            "symbol": "us/west",
+            "bid": 100.0,
+            "ask": 101.0,
+            "details": {"created_by": "test"},
+        }
+    ]
+    df = pa.Table.from_pylist(data)
+    table.append(df)
+
+    create_clickhouse_iceberg_database(started_cluster, node, CATALOG_NAME)
+    assert 1 == int(node.query(f"SELECT count() FROM {CATALOG_NAME}.`{root_namespace}.{table_name}`"))
+    assert "us/west" in node.query(f"SELECT symbol FROM {CATALOG_NAME}.`{root_namespace}.{table_name}`")
+
+
 def test_cluster_select(started_cluster):
     node1 = started_cluster.instances["node1"]
     node2 = started_cluster.instances["node2"]
@@ -665,7 +711,7 @@ def test_cluster_select(started_cluster):
         assert len(cluster_secondary_queries) == 1
 
     assert node2.query(f"SELECT * FROM {CATALOG_NAME}.`{root_namespace}.{table_name}`", settings={"parallel_replicas_for_cluster_engines":1, 'enable_parallel_replicas': 2, 'cluster_for_parallel_replicas': 'cluster_simple', 'parallel_replicas_for_cluster_engines' : 1}) == 'pablo\n'
-    
+
 def test_not_specified_catalog_type(started_cluster):
     node = started_cluster.instances["node1"]
     settings = {