Revert "tmp"

This reverts commit ee9e4e5.

Author: arthurpassos

src/Storages/VirtualColumnUtils.cpp

@@ -2,7 +2,6 @@
 #include <stack>
 
 #include <Storages/VirtualColumnUtils.h>
-#include <absl/container/flat_hash_set.h>
 #include "Formats/NumpyDataTypes.h"
 
 #include <Core/NamesAndTypes.h>
@@ -154,27 +153,6 @@ static auto makeExtractor()
     return KeyValuePairExtractorBuilder().withItemDelimiters({'/'}).withKeyValueDelimiter('=').buildWithReferenceMap();
 }
 
-HivePartitioningKeysAndValues parseHivePartitioningKeysAndValuesRegex(const String & path)
-{
-    const static RE2 pattern_re("([^/]+)=([^/]*)/");
-    re2::StringPiece input_piece(path);
-
-    HivePartitioningKeysAndValues result;
-    std::string_view key;
-    std::string_view value;
-
-    while (RE2::FindAndConsume(&input_piece, pattern_re, &key, &value))
-    {
-        auto it = result.find(key);
-        if (it != result.end() && it->second != value)
-            throw Exception(ErrorCodes::INCORRECT_DATA, "Path '{}' to file with enabled hive-style partitioning contains duplicated partition key {} with different values, only unique keys are allowed", path, key);
-
-        auto col_name = key;
-        result[col_name] = value;
-    }
-    return result;
-}
-
 HivePartitioningKeysAndValues parseHivePartitioningKeysAndValues(const String & path)
 {
     static auto extractor = makeExtractor();
@@ -205,45 +183,6 @@ HivePartitioningKeysAndValues parseHivePartitioningKeysAndValues(const String &
     return key_values;
 }
 
-std::pair<ColumnPtr, ColumnPtr> parseHivePartitioningKeysAndValuesOldExtractkv(const String & path)
-{
-    static auto extractor = KeyValuePairExtractorBuilder().withItemDelimiters({'/'}).withKeyValueDelimiter('=').buildWithoutEscaping();
-
-    auto keys = ColumnString::create();
-    auto values = ColumnString::create();
-
-    // cutting the filename to prevent malformed filenames that contain key-value-pairs from being extracted
-    // not sure if we actually need to do that, but just in case. Plus, the previous regex impl took care of it
-    const auto last_slash_pos = path.find_last_of('/');
-
-    if (last_slash_pos == std::string::npos)
-    {
-        // nothing to extract, there is no path, just a filename
-        return std::make_pair(std::move(keys), std::move(values));
-    }
-
-    std::string_view path_without_filename(path.data(), last_slash_pos);
-
-    extractor.extract(path_without_filename, keys, values);
-
-    keys->validate();
-    values->validate();
-
-    absl::flat_hash_set<StringRef> check_for_duplicates_set;
-
-    for (auto i = 0u; i < keys->size(); i++)
-    {
-        auto [_, inserted] = check_for_duplicates_set.insert(keys->getDataAt(i));
-
-        if (!inserted)
-        {
-            throw Exception(ErrorCodes::INCORRECT_DATA, "Path '{}' to file with enabled hive-style partitioning contains duplicated partition key {} with different values, only unique keys are allowed", path, keys->getDataAt(i).toString());
-        }
-    }
-
-    return std::make_pair(std::move(keys), std::move(values));
-}
-
 VirtualColumnsDescription getVirtualsForFileLikeStorage(ColumnsDescription & storage_columns, const ContextPtr & context, const std::string & path, std::optional<FormatSettings> format_settings_)
 {
     VirtualColumnsDescription desc;

src/Storages/VirtualColumnUtils.h

@@ -112,9 +112,7 @@ void addRequestedFileLikeStorageVirtualsToChunk(
 
 using HivePartitioningKeysAndValues = absl::flat_hash_map<std::string_view, std::string_view>;
 
-HivePartitioningKeysAndValues parseHivePartitioningKeysAndValuesRegex(const String & path);
 HivePartitioningKeysAndValues parseHivePartitioningKeysAndValues(const String & path);
-std::pair<ColumnPtr, ColumnPtr> parseHivePartitioningKeysAndValuesOldExtractkv(const String & path);
 
 }
 

src/Storages/tests/gtest_virtual_column_utils.cpp

@@ -85,64 +85,3 @@ TEST(VirtualColumnUtils, getVirtualsForFileLikeStorageEmptyValue)
     ASSERT_TRUE(res.has("year"));
     ASSERT_TRUE(res.has("country"));
 }
-
-static std::vector<std::string> test_paths = {
-    "/some/folder/key1=val1/key2=val2/file1.txt",
-    "/data/keyA=valA/keyB=valB/keyC=valC/file2.txt",
-    "/another/dir/x=1/y=2/z=3/file3.txt",
-    "/tiny/path/a=b/file4.txt",
-    "/yet/another/path/k1=v1/k2=v2/k3=v3/k4=v4/k5=v5/"
-};
-
-TEST(VirtualColumnUtils, BenchmarkRegexParser)
-{
-    static constexpr int iterations = 100000;
-
-    auto start = std::chrono::steady_clock::now();
-
-    for (int i = 0; i < iterations; ++i)
-    {
-        const auto & path = test_paths[i % 5];
-        auto result = parseHivePartitioningKeysAndValuesRegex(path);
-        ASSERT_TRUE(!result.empty());
-    }
-
-    auto end = std::chrono::steady_clock::now();
-    auto duration_ms = std::chrono::duration_cast<std::chrono::milliseconds>(end - start).count();
-
-    std::cout << "[BenchmarkRegexParser] "
-              << iterations << " iterations across 5 paths took "
-              << duration_ms << " ms\n";
-
-    auto start_extractkv = std::chrono::steady_clock::now();
-
-    for (int i = 0; i < iterations; ++i)
-    {
-        const auto & path = test_paths[i % 5];
-        auto result = parseHivePartitioningKeysAndValues(path);
-        ASSERT_TRUE(!result.empty());
-    }
-
-    auto end_extractkv = std::chrono::steady_clock::now();
-    auto duration_ms_extractkv = std::chrono::duration_cast<std::chrono::milliseconds>(end_extractkv - start_extractkv).count();
-
-    std::cout << "[BenchmarkExtractkvParser] "
-              << iterations << " iterations across 5 paths took "
-              << duration_ms_extractkv << " ms\n";
-
-    auto start_extractkv_old = std::chrono::steady_clock::now();
-
-    for (int i = 0; i < iterations; ++i)
-    {
-        const auto & path = test_paths[i % 5];
-        auto result = parseHivePartitioningKeysAndValuesOldExtractkv(path);
-        ASSERT_TRUE(!result.first->empty());
-    }
-
-    auto end_extractkv_old = std::chrono::steady_clock::now();
-    auto duration_ms_extractkv_old = std::chrono::duration_cast<std::chrono::milliseconds>(end_extractkv_old - start_extractkv_old).count();
-
-    std::cout << "[BenchmarkExtractkvOldParser] "
-              << iterations << " iterations across 5 paths took "
-              << duration_ms_extractkv_old << " ms\n";
-}

tests/queries/0_stateless/03203_hive_style_partitioning.reference

@@ -95,3 +95,24 @@ test/hive_partitioning/column0=Arthur/column1=/sample.parquet
 test/hive_partitioning/column0=Arthur/column1=ABC/sample.parquet
 test/hive_partitioning/column0=Arthur/column1=/sample.parquet
 OK
+TESTING THE S3CLUSTER PARTITIONING
+ last	Elizabeth
+Frank	Elizabeth
+Moreno	Elizabeth
+Guzman	Elizabeth
+Stephens	Elizabeth
+Franklin	Elizabeth
+Gibson	Elizabeth
+Greer	Elizabeth
+Delgado	Elizabeth
+Cross	Elizabeth
+ last	Elizabeth
+Frank	Elizabeth
+Moreno	Elizabeth
+Guzman	Elizabeth
+Stephens	Elizabeth
+Franklin	Elizabeth
+Gibson	Elizabeth
+Greer	Elizabeth
+Delgado	Elizabeth
+Cross	Elizabeth

tests/queries/0_stateless/03203_hive_style_partitioning.sh

@@ -79,3 +79,13 @@ set use_hive_partitioning = 0;
 
 SELECT *, _column0 FROM s3('http://localhost:11111/test/hive_partitioning/column0=Elizabeth/sample.parquet') LIMIT 10;
 """ 2>&1 | grep -F -q "UNKNOWN_IDENTIFIER" && echo "OK" || echo "FAIL";
+
+$CLICKHOUSE_LOCAL -q "SELECT 'TESTING THE S3CLUSTER PARTITIONING'"
+
+$CLICKHOUSE_CLIENT -q """
+set use_hive_partitioning = 1;
+
+SELECT *, column0 FROM s3Cluster(test_cluster_one_shard_three_replicas_localhost, 'http://localhost:11111/test/hive_partitioning/column0=Elizabeth/sample.parquet') LIMIT 10;
+
+SELECT *, column0 FROM s3Cluster(test_cluster_one_shard_three_replicas_localhost, 'http://localhost:11111/test/hive_partitioning/column0=Elizabeth/sample.parquet') WHERE column0 = 'Elizabeth' LIMIT 10;
+"""