Merge pull request ClickHouse#87742 from ClickHouse/backport/25.8/84020

Backport ClickHouse#84020 to 25.8: Fix geoparquet types breaking client-server protocol

Author: clickhouse-gh[bot]

src/Formats/SchemaInferenceUtils.cpp

@@ -1639,6 +1639,11 @@ DataTypePtr makeNullableRecursively(DataTypePtr type, const FormatSettings & set
     if (which.isNullable())
         return type;
 
+    /// Leave named compound types unchanged.
+    /// E.g. don't turn `Point` into `Tuple(Nullable(Float64), Nullable(Float64))`.
+    if (type->hasCustomName())
+        return makeNullableSafe(type);
+
     if (which.isArray())
     {
         const auto * array_type = assert_cast<const DataTypeArray *>(type.get());
@@ -1710,25 +1715,8 @@ DataTypePtr makeNullableRecursively(DataTypePtr type, const FormatSettings & set
 NamesAndTypesList getNamesAndRecursivelyNullableTypes(const Block & header, const FormatSettings & settings)
 {
     NamesAndTypesList result;
-
-    std::unordered_map<String, DataTypeCustomDescPtr> custom_descs;
-    const auto & prev_schema = header.getNamesAndTypesList();
-    for (const auto & [name, type] : prev_schema)
-        if (type->hasCustomName() && (type->getTypeId() == TypeIndex::Tuple || type->getTypeId() == TypeIndex::Array))
-            custom_descs[name] = std::make_unique<DataTypeCustomDesc>(std::make_unique<DataTypeCustomFixedName>(type->getCustomName()->getName()));
-
     for (auto & [name, type] : header.getNamesAndTypesList())
         result.emplace_back(name, makeNullableRecursively(type, settings));
-
-    for (auto & [name, type] : result)
-    {
-        auto it = custom_descs.find(name);
-        if (it != custom_descs.end())
-        {
-            type->setCustomization(std::move(it->second));
-        }
-    }
-
     return result;
 }
 

src/Formats/SchemaInferenceUtils.h

@@ -101,11 +101,15 @@ void transformFinalInferredJSONTypeIfNeeded(DataTypePtr & data_type, const Forma
 void transformInferredJSONTypesFromDifferentFilesIfNeeded(DataTypePtr & first, DataTypePtr & second, const FormatSettings & settings);
 
 /// Make type Nullable recursively:
-/// - Type -> Nullable(type)
-/// - Array(Type) -> Array(Nullable(Type))
-/// - Tuple(Type1, ..., TypeN) -> Tuple(Nullable(Type1), ..., Nullable(TypeN))
-/// - Map(KeyType, ValueType) -> Map(KeyType, Nullable(ValueType))
-/// - LowCardinality(Type) -> LowCardinality(Nullable(Type))
+///  - Type -> Nullable(type)
+///  - Array(Type) -> Array(Nullable(Type))
+///  - Tuple(Type1, ..., TypeN) -> Tuple(Nullable(Type1), ..., Nullable(TypeN))
+///  - Map(KeyType, ValueType) -> Map(KeyType, Nullable(ValueType))
+///  - LowCardinality(Type) -> LowCardinality(Nullable(Type))
+/// Does not recurse into types with custom name.
+/// E.g. type `Point` (aka `Tuple(Float64, Float64)`) stays unchanged as `Point`, it does not become
+/// `Tuple(Nullable(Float64), Nullable(Float64))`.
+/// But `Bool` becomes `Nullable(Bool)`.
 DataTypePtr makeNullableRecursively(DataTypePtr type, const FormatSettings & settings);
 
 /// Call makeNullableRecursively for all types

tests/queries/0_stateless/03571_geoparquet_nullable_bug.reference

@@ -0,0 +1,2 @@
+(10,20)	[(0,0),(10,0),(10,10),(0,10)]	[[(20,20),(50,20),(50,50),(20,50)],[(30,30),(50,50),(50,30)]]	[[(0,0),(10,0),(10,10),(0,10)],[(1,1),(2,2),(3,3)]]	[[[(0,0),(10,0),(10,10),(0,10)]],[[(20,20),(50,20),(50,50),(20,50)],[(30,30),(50,50),(50,30)]]]
+(30,40)	[(1,1),(11,1),(11,11),(1,11)]	[[(21,21),(51,21),(51,51),(21,51)],[(31,31),(51,51),(51,31)]]	[[(1,1),(11,1),(11,11),(1,11)],[(2,2),(3,3),(4,4)]]	[[[(1,1),(11,1),(11,11),(1,11)]],[[(21,21),(51,21),(51,51),(21,51)],[(31,31),(51,51),(51,31)]]]

tests/queries/0_stateless/03571_geoparquet_nullable_bug.sh

@@ -0,0 +1,24 @@
+#!/usr/bin/env bash
+# Tags: no-fasttest
+
+CURDIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)
+# shellcheck source=../shell_config.sh
+. "$CURDIR"/../shell_config.sh
+
+$CLICKHOUSE_CLIENT --query="
+    insert into function file('${CLICKHOUSE_TEST_UNIQUE_NAME}.parquet', Parquet, 'point Point, linestring LineString, polygon Polygon, multilinestring MultiLineString, multipolygon MultiPolygon') values (
+            (10, 20),
+            [(0, 0), (10, 0), (10, 10), (0, 10)],
+            [[(20, 20), (50, 20), (50, 50), (20, 50)], [(30, 30), (50, 50), (50, 30)]],
+            [[(0, 0), (10, 0), (10, 10), (0, 10)], [(1, 1), (2, 2), (3, 3)]],
+            [[[(0, 0), (10, 0), (10, 10), (0, 10)]], [[(20, 20), (50, 20), (50, 50), (20, 50)],[(30, 30), (50, 50), (50, 30)]]]
+        ), (
+            (30, 40),
+            [(1, 1), (11, 1), (11, 11), (1, 11)],
+            [[(21, 21), (51, 21), (51, 51), (21, 51)], [(31, 31), (51, 51), (51, 31)]],
+            [[(1, 1), (11, 1), (11, 11), (1, 11)], [(2, 2), (3, 3), (4, 4)]],
+            [[[(1, 1), (11, 1), (11, 11), (1, 11)]], [[(21, 21), (51, 21), (51, 51), (21, 51)],[(31, 31), (51, 51), (51, 31)]]]
+        );
+    select * from file('${CLICKHOUSE_TEST_UNIQUE_NAME}.parquet');"
+
+rm "${CLICKHOUSE_USER_FILES}/${CLICKHOUSE_TEST_UNIQUE_NAME}.parquet"