clickhouse-native 0.9.0 → 0.11.0

data/ext/clickhouse_native/vendor/clickhouse-cpp/clickhouse/columns/tuple.h

@@ -13,6 +13,8 @@ namespace clickhouse {
 class ColumnTuple : public Column {
 public:
     ColumnTuple(const std::vector<ColumnRef>& columns);
+    ColumnTuple(const std::vector<ColumnRef>& columns,
+                std::vector<std::string> names);
 
     /// Returns count of columns in the tuple.
     size_t TupleSize() const;
@@ -69,12 +71,22 @@ public:
     ColumnTupleT(std::tuple<std::shared_ptr<Columns>...> columns)
         : ColumnTuple(TupleToVector(columns)), typed_columns_(std::move(columns)) {}
 
+    ColumnTupleT(std::tuple<std::shared_ptr<Columns>...> columns, std::vector<std::string> names)
+        : ColumnTuple(TupleToVector(columns), std::move(names)), typed_columns_(std::move(columns)) {}
+
     ColumnTupleT(std::vector<ColumnRef> columns)
         : ColumnTuple(columns), typed_columns_(VectorToTuple(std::move(columns))) {}
 
+    ColumnTupleT(std::vector<ColumnRef> columns, std::vector<std::string> names)
+        : ColumnTuple(columns, std::move(names)), typed_columns_(VectorToTuple(std::move(columns))) {}
+
     ColumnTupleT(const std::initializer_list<ColumnRef> columns)
         : ColumnTuple(columns), typed_columns_(VectorToTuple(std::move(columns))) {}
 
+    ColumnTupleT(std::initializer_list<ColumnRef> columns, std::vector<std::string> names)
+        : ColumnTuple(std::vector<ColumnRef>(columns), std::move(names))
+        , typed_columns_(VectorToTuple(std::vector<ColumnRef>(columns))) {}
+
     inline ValueType At(size_t index) const { return GetTupleOfValues(index); }
 
     inline ValueType operator[](size_t index) const { return GetTupleOfValues(index); }
@@ -99,7 +111,8 @@ public:
         if (col.TupleSize() != std::tuple_size_v<TupleOfColumns>) {
             throw ValidationError("Can't wrap from " + col.GetType().GetName());
         }
-        return std::make_shared<ColumnTupleT<Columns...>>(VectorToTuple(std::move(col)));
+        auto names = col.Type()->As<TupleType>()->GetItemNames();
+        return std::make_shared<ColumnTupleT<Columns...>>(VectorToTuple(std::move(col)), std::move(names));
     }
 
     static auto Wrap(Column&& col) { return Wrap(std::move(dynamic_cast<ColumnTuple&&>(col))); }

data/ext/clickhouse_native/vendor/clickhouse-cpp/clickhouse/query.h

@@ -229,8 +229,8 @@ private:
     }
 
 private:
-    const std::string query_;
-    const std::string query_id_;
+    std::string query_;
+    std::string query_id_;
     std::optional<open_telemetry::TracingContext> tracing_context_;
     QuerySettings query_settings_;
     QueryParams query_params_;

data/ext/clickhouse_native/vendor/clickhouse-cpp/clickhouse/server_exception.h

@@ -1,7 +1,6 @@
 #pragma once
 
 #include <string>
-#include <memory>
 
 namespace clickhouse {
 struct Exception {
@@ -9,8 +8,6 @@ struct Exception {
     std::string name;
     std::string display_text;
     std::string stack_trace;
-    /// Pointer to nested exception.
-    std::unique_ptr<Exception> nested;
 };
 
 }

data/ext/clickhouse_native/vendor/clickhouse-cpp/clickhouse/types/type_parser.cpp

@@ -22,7 +22,9 @@ bool TypeAst::operator==(const TypeAst & other) const {
     return meta == other.meta
         && code == other.code
         && name == other.name
+        && element_name == other.element_name
         && value == other.value
+        && value_string == other.value_string
         && std::equal(elements.begin(), elements.end(), other.elements.begin(), other.elements.end());
 }
 
@@ -32,7 +34,11 @@ static const std::unordered_map<std::string, Type::Code> kTypeCode = {
     { "Int16",       Type::Int16 },
     { "Int32",       Type::Int32 },
     { "Int64",       Type::Int64 },
+#if CH_MAP_BOOL_TO_UINT8
     { "Bool",        Type::UInt8 },
+#else
+    { "Bool",        Type::Bool },
+#endif
     { "UInt8",       Type::UInt8 },
     { "UInt16",      Type::UInt16 },
     { "UInt32",      Type::UInt32 },
@@ -67,6 +73,7 @@ static const std::unordered_map<std::string, Type::Code> kTypeCode = {
     { "MultiPolygon", Type::MultiPolygon },
     { "Time",        Type::Time },
     { "Time64",      Type::Time64 },
+    { "JSON",        Type::JSON },
 };
 
 template <typename L, typename R>
@@ -166,7 +173,14 @@ bool TypeParser::Parse(TypeAst* type) {
                 type_->code = Type::String;
                 break;
             }
+            case Token::QuotedIdentifier:
             case Token::Name:
+                if (!type_->name.empty()) {
+                    // A second Name token on the same element means the
+                    // previous one was a field name in a named-tuple element
+                    // (e.g. "a" in "Tuple(a Int32, …)").
+                    type_->element_name = std::move(type_->name);
+                }
                 type_->meta = GetTypeMeta(token.value);
                 type_->name = token.value.to_string();
                 type_->code = GetTypeCode(type_->name);
@@ -247,6 +261,35 @@ TypeParser::Token TypeParser::NextToken() {
                 }
                 return Token{Token::QuotedString, StringView(cur_++, 1)};
             }
+            case '"':
+            case '`':
+            {
+                const auto quote = *cur_;
+                ++cur_;
+                // Two escape forms are recognised, both quote-specific (e.g.
+                // inside a backtick-quoted identifier only backtick escapes
+                // apply; a doubled double-quote is treated as two literals):
+                //   \q  – backslash followed by the opening quote character
+                //   qq  – two consecutive opening quote characters
+                scratch_.clear();
+                for (; cur_ < end_; ++cur_) {
+                    if (*cur_ == '\\' && cur_ + 1 < end_ && *(cur_ + 1) == quote) {
+                        scratch_ += quote;
+                        ++cur_;
+                    } else if (*cur_ == quote) {
+                        if (cur_ + 1 < end_ && *(cur_ + 1) == quote) {
+                            scratch_ += quote;
+                            ++cur_;
+                        } else {
+                            ++cur_;
+                            return Token{Token::QuotedIdentifier, StringView{scratch_}};
+                        }
+                    } else {
+                        scratch_ += *cur_;
+                    }
+                }
+                return Token{Token::Invalid, StringView()};
+            }
 
             default: {
                 const char* st = cur_;

data/ext/clickhouse_native/vendor/clickhouse-cpp/clickhouse/types/type_parser.h

@@ -31,6 +31,9 @@ struct TypeAst {
     /// Type's name.
     /// Need to cache TypeAst, so can't use StringView for name.
     std::string name;
+    /// Name of this element inside its parent (e.g. field name inside a named
+    /// Tuple). Empty for unnamed elements.
+    std::string element_name;
     /// Value associated with the node,
     /// used for fixed-width types and enum values.
     int64_t value = 0;
@@ -59,6 +62,7 @@ class TypeParser {
             RPar,
             Comma,
             QuotedString, // string with quotation marks included
+            QuotedIdentifier,
             EOS,
         };
 
@@ -81,6 +85,11 @@ private:
 
     TypeAst* type_;
     std::stack<TypeAst*> open_elements_;
+    // Backing storage for unescaped QuotedIdentifier token values. When a
+    // quoted identifier contains escape sequences the unescaped content is
+    // written here and the returned StringView points into this string.
+    // Valid only until the next NextToken() call.
+    std::string scratch_;
 };
 
 

data/ext/clickhouse_native/vendor/clickhouse-cpp/clickhouse/types/types.cpp

@@ -54,6 +54,8 @@ const char* Type::TypeName(Type::Code code) {
         case Type::Code::MultiPolygon:   return "MultiPolygon";
         case Type::Code::Time:           return "Time";
         case Type::Code::Time64:         return "Time64";
+        case Type::Code::JSON:           return "JSON";
+        case Type::Code::Bool:           return "Bool";
     }
 
     return "Unknown type";
@@ -85,6 +87,8 @@ std::string Type::GetName() const {
         case Ring:
         case Polygon:
         case MultiPolygon:
+        case JSON:
+        case Bool:
             return TypeName(code_);
         case Time64:
             return As<Time64Type>()->GetName();
@@ -138,6 +142,7 @@ uint64_t Type::GetTypeUniqueId() const {
         case Float32:
         case Float64:
         case String:
+        case JSON:
         case IPv4:
         case IPv6:
         case Date:
@@ -146,6 +151,7 @@ uint64_t Type::GetTypeUniqueId() const {
         case Ring:
         case Polygon:
         case MultiPolygon:
+        case Bool:
             // For simple types, unique ID is the same as Type::Code
             return code_;
 
@@ -239,8 +245,9 @@ TypeRef Type::CreateString(size_t n) {
     return TypeRef(new FixedStringType(n));
 }
 
-TypeRef Type::CreateTuple(const std::vector<TypeRef>& item_types) {
-    return TypeRef(new TupleType(item_types));
+TypeRef Type::CreateTuple(const std::vector<TypeRef>& item_types,
+                          std::vector<std::string> item_names) {
+    return TypeRef(new TupleType(item_types, std::move(item_names)));
 }
 
 TypeRef Type::CreateEnum8(const std::vector<EnumItem>& enum_items) {
@@ -279,6 +286,10 @@ TypeRef Type::CreateMultiPolygon() {
     return TypeRef(new Type(Type::MultiPolygon));
 }
 
+TypeRef Type::CreateJSON() {
+    return TypeRef(new Type(Type::JSON));
+}
+
 /// class ArrayType
 
 ArrayType::ArrayType(TypeRef item_type) : Type(Array), item_type_(item_type) {
@@ -442,9 +453,17 @@ FixedStringType::FixedStringType(size_t n) : Type(FixedString), size_(n) {
 NullableType::NullableType(TypeRef nested_type) : Type(Nullable), nested_type_(nested_type) {
 }
 
-/// class TupleType
-
-TupleType::TupleType(const std::vector<TypeRef>& item_types) : Type(Tuple), item_types_(item_types) {
+TupleType::TupleType(const std::vector<TypeRef>& item_types,
+                     std::vector<std::string> item_names)
+    : Type(Tuple), item_types_(item_types), item_names_(std::move(item_names)) {
+    if (!item_names_.empty() && item_names_.size() != item_types_.size()) {
+        throw ValidationError("Tuple field names count doesn't match tuple element count");
+    }
+    for (const auto& item_name : item_names_) {
+        if (item_name.empty()) {
+            throw ValidationError("Tuple field names can't be empty");
+        }
+    }
 }
 
 /// class LowCardinalityType
@@ -454,15 +473,46 @@ LowCardinalityType::LowCardinalityType(TypeRef nested_type) : Type(LowCardinalit
 LowCardinalityType::~LowCardinalityType() {
 }
 
-std::string TupleType::GetName() const {
-    std::string result("Tuple(");
+// Checks if `name` is a valid plain identifier (must not be quoted).
+// The condition for this is a match against `^[a-zA-Z_][0-9a-zA-Z_]*$`
+static bool IsPlainIdentifier(const std::string& name) {
+    if (name.empty()) return false;
+    auto is_alpha_or_under = [](char c) { return (c >= 'a' && c <= 'z') || (c >= 'A' && c <= 'Z') || c == '_'; };
+    auto is_alnum_or_under = [&is_alpha_or_under](char c) { return is_alpha_or_under(c) || (c >= '0' && c <= '9'); };
+    if (!is_alpha_or_under(name[0])) return false;
+    for (size_t i = 1; i < name.size(); ++i)
+        if (!is_alnum_or_under(name[i])) return false;
+    return true;
+}
 
-    if (!item_types_.empty()) {
-        result += item_types_[0]->GetName();
+// Appends a fieldname, potentially quoting it and escaping backticks.
+static void AppendFieldname(const std::string& name, std::string& out) {
+    if (IsPlainIdentifier(name)) {
+        out += name;
+        return;
+    }
+    out += '`';
+    for (char c : name) {
+        if (c == '`')
+            out += "``";
+        else
+            out += c;
     }
+    out += '`';
+}
+
+std::string TupleType::GetName() const {
+    std::string result("Tuple(");
+    bool has_complete_names = !item_names_.empty();
 
-    for (size_t i = 1; i < item_types_.size(); ++i) {
-        result += ", " + item_types_[i]->GetName();
+    for (size_t i = 0; i < item_types_.size(); ++i) {
+        if (i > 0)
+            result += ", ";
+        if (has_complete_names) {
+            AppendFieldname(item_names_[i], result);
+            result += ' ';
+        }
+        result += item_types_[i]->GetName();
     }
 
     result += ")";

data/ext/clickhouse_native/vendor/clickhouse-cpp/clickhouse/types/types.h

@@ -59,6 +59,8 @@ public:
         MultiPolygon,
         Time,
         Time64,
+        JSON,
+        Bool,
     };
 
     using EnumItem = std::pair<std::string /* name */, int16_t /* value */>;
@@ -124,7 +126,8 @@ public:
 
     static TypeRef CreateString(size_t n);
 
-    static TypeRef CreateTuple(const std::vector<TypeRef>& item_types);
+    static TypeRef CreateTuple(const std::vector<TypeRef>& item_types,
+                               std::vector<std::string> item_names = {});
 
     static TypeRef CreateEnum8(const std::vector<EnumItem>& enum_items);
 
@@ -148,6 +151,8 @@ public:
 
     static TypeRef CreateTime64(size_t precision);
 
+    static TypeRef CreateJSON();
+
 private:
     uint64_t GetTypeUniqueId() const;
 
@@ -292,15 +297,21 @@ private:
 
 class TupleType : public Type {
 public:
-    explicit TupleType(const std::vector<TypeRef>& item_types);
+    explicit TupleType(const std::vector<TypeRef>& item_types,
+                       std::vector<std::string> item_names = {});
 
     std::string GetName() const;
 
     /// Type of nested Tuple element type.
     std::vector<TypeRef> GetTupleType() const { return item_types_; }
 
+    /// Field names for named tuples. Same length as GetTupleType() when
+    /// populated, or empty when the tuple has no field names.
+    const std::vector<std::string>& GetItemNames() const { return item_names_; }
+
 private:
     std::vector<TypeRef> item_types_;
+    std::vector<std::string> item_names_;
 };
 
 class LowCardinalityType : public Type {
@@ -384,6 +395,11 @@ inline TypeRef Type::CreateSimple<uint64_t>() {
     return TypeRef(new Type(UInt64));
 }
 
+template <>
+inline TypeRef Type::CreateSimple<bool>() {
+    return TypeRef(new Type(Bool));
+}
+
 template <>
 inline TypeRef Type::CreateSimple<float>() {
     return TypeRef(new Type(Float32));

data/ext/clickhouse_native/vendor/clickhouse-cpp/clickhouse/version.h

@@ -2,7 +2,7 @@
 
 #define CLICKHOUSE_CPP_VERSION_MAJOR 2
 #define CLICKHOUSE_CPP_VERSION_MINOR 6
-#define CLICKHOUSE_CPP_VERSION_PATCH 1
+#define CLICKHOUSE_CPP_VERSION_PATCH 2
 
 #define CLICKHOUSE_CPP_VERSION_BUILD 0
 

data/lib/clickhouse_native/client.rb

@@ -2,7 +2,7 @@
 
 module ClickhouseNative
   class Client
-    attr_reader :host, :port, :database
+    attr_reader :host, :port, :database, :ping_before_query, :tcp_keepalive, :retry_timeout
 
     def describe_table(table, db_name: nil)
       fq = db_name ? "#{db_name}.#{table}" : table

data/lib/clickhouse_native/pool.rb

@@ -8,16 +8,17 @@ module ClickhouseNative
 
     def initialize(host:, port:, database: "default", user: "default", password: "",
                    compression: :none, logger: nil, settings: {},
-                   pool_size: 5, pool_timeout: 5)
+                   pool_size: 5, pool_timeout: 5,
+                   ping_before_query: true, tcp_keepalive: true, retry_timeout: 1)
       @host = host
       @port = port
       @database = database
-      client_kwargs = { host:, port:, database:, user:, password:, compression:, logger: }
-      @set_sql = settings_sql(settings)
+      client_kwargs = {
+        host:, port:, database:, user:, password:, compression:, logger:, settings:,
+        ping_before_query:, tcp_keepalive:, retry_timeout:
+      }
       @pool = ConnectionPool.new(size: pool_size, timeout: pool_timeout) do
-        client = Client.new(**client_kwargs)
-        client.execute(@set_sql) if @set_sql
-        client
+        Client.new(**client_kwargs)
       end
     end
 
@@ -25,18 +26,24 @@ module ClickhouseNative
     # path leaves the socket in an unknown state. The C++ binding issues
     # ResetConnection, but a subsequent send can still surface buffered
     # protocol errors from the prior aborted operation — those get
-    # attributed to whatever SQL we tried next (e.g. the SET reapplying
-    # session settings), producing misleading log lines and re-raises in
-    # unrelated code. A fresh socket + handshake is cheap relative to
-    # debugging that.
+    # attributed to whatever SQL we tried next, producing misleading log
+    # lines and re-raises in unrelated code. A fresh socket + handshake is
+    # cheap relative to debugging that.
     #
     # ConnectionError gets one automatic retry: pooled connections that
     # have been idle long enough for the server / an LB to FIN them
-    # surface as "closed" on the very next recv (errno 0, message
-    # "closed: Success"). Discarding and re-checking out lands a fresh
-    # socket and the operation succeeds. The retry only triggers when
-    # the dead-connection error fired before any data was sent, so
-    # write operations don't risk double-execution from this path.
+    # surface as "closed" on the very next recv (errno is whatever stale
+    # value was left in the thread — "closed: Success", "closed: Operation
+    # now in progress", etc. all mean the same recv()==0). Discarding and
+    # re-checking out lands a fresh socket and the operation succeeds. The
+    # retry only triggers when the dead-connection error fired before any
+    # data was sent, so write operations don't risk double-execution.
+    #
+    # This is the backstop: with ping_before_query on (the default) the
+    # driver already pings and transparently reconnects a dead socket
+    # before running the query, so most stale connections never surface as
+    # a ConnectionError here at all. This retry still covers the residual
+    # race (socket dies between the ping and the query).
     def with
       attempts = 0
       begin
@@ -84,21 +91,5 @@ module ClickhouseNative
     def describe_table(table, db_name: nil)
       with { |c| c.describe_table(table, db_name:) }
     end
-
-    private
-
-    # Render a `SET key1 = val1, key2 = val2` statement once at pool setup
-    # so every checked-out connection starts with the same session
-    # settings. Matches how the HTTP driver injected global_params per
-    # request. Values: Integer / Float render bare; anything else is
-    # quoted as a SQL string literal.
-    def settings_sql(settings)
-      return nil if settings.nil? || settings.empty?
-      parts = settings.map do |k, v|
-        literal = v.is_a?(Numeric) ? v.to_s : "'#{v.to_s.gsub("'", "''")}'"
-        "#{k} = #{literal}"
-      end
-      "SET #{parts.join(', ')}"
-    end
   end
 end

data/lib/clickhouse_native/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module ClickhouseNative
-  VERSION = "0.9.0"
+  VERSION = "0.11.0"
 end

data/lib/clickhouse_native.rb

@@ -3,6 +3,7 @@
 require "time"
 require "date"
 require "bigdecimal"
+require "json"
 require "clickhouse_native/version"
 require "clickhouse_native/errors"
 

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: clickhouse-native
 version: !ruby/object:Gem::Version
-  version: 0.9.0
+  version: 0.11.0
 platform: ruby
 authors:
 - Yuri Smirnov
@@ -35,19 +35,29 @@ files:
 - ext/clickhouse_native/client.cpp
 - ext/clickhouse_native/extconf.rb
 - ext/clickhouse_native/patches/0001-preserve-declared-column-type.patch
+- ext/clickhouse_native/patches/0002-carry-settings-into-insert.patch
 - ext/clickhouse_native/vendor/clickhouse-cpp/.clang-format
 - ext/clickhouse_native/vendor/clickhouse-cpp/.git
 - ext/clickhouse_native/vendor/clickhouse-cpp/.gitattributes
 - ext/clickhouse_native/vendor/clickhouse-cpp/.github/CODEOWNERS
+- ext/clickhouse_native/vendor/clickhouse-cpp/.github/workflows/bazel.yml
+- ext/clickhouse_native/vendor/clickhouse-cpp/.github/workflows/cross-repo-bug-relay.yml
 - ext/clickhouse_native/vendor/clickhouse-cpp/.github/workflows/linux.yml
 - ext/clickhouse_native/vendor/clickhouse-cpp/.github/workflows/macos.yml
 - ext/clickhouse_native/vendor/clickhouse-cpp/.github/workflows/windows_mingw.yml
 - ext/clickhouse_native/vendor/clickhouse-cpp/.github/workflows/windows_msvc.yml
 - ext/clickhouse_native/vendor/clickhouse-cpp/.gitignore
 - ext/clickhouse_native/vendor/clickhouse-cpp/.travis.yml
+- ext/clickhouse_native/vendor/clickhouse-cpp/AI_POLICY.md
+- ext/clickhouse_native/vendor/clickhouse-cpp/BUILD.bazel
 - ext/clickhouse_native/vendor/clickhouse-cpp/CMakeLists.txt
 - ext/clickhouse_native/vendor/clickhouse-cpp/LICENSE
+- ext/clickhouse_native/vendor/clickhouse-cpp/MODULE.bazel
+- ext/clickhouse_native/vendor/clickhouse-cpp/MODULE.bazel.lock
 - ext/clickhouse_native/vendor/clickhouse-cpp/README.md
+- ext/clickhouse_native/vendor/clickhouse-cpp/ci/docker-compose.yml
+- ext/clickhouse_native/vendor/clickhouse-cpp/ci/docker-compose/config.xml
+- ext/clickhouse_native/vendor/clickhouse-cpp/ci/docker-compose/users.xml
 - ext/clickhouse_native/vendor/clickhouse-cpp/clickhouse/CMakeLists.txt
 - ext/clickhouse_native/vendor/clickhouse-cpp/clickhouse/base/buffer.h
 - ext/clickhouse_native/vendor/clickhouse-cpp/clickhouse/base/compressed.cpp
@@ -78,6 +88,8 @@ files:
 - ext/clickhouse_native/vendor/clickhouse-cpp/clickhouse/client.h
 - ext/clickhouse_native/vendor/clickhouse-cpp/clickhouse/columns/array.cpp
 - ext/clickhouse_native/vendor/clickhouse-cpp/clickhouse/columns/array.h
+- ext/clickhouse_native/vendor/clickhouse-cpp/clickhouse/columns/bool.cpp
+- ext/clickhouse_native/vendor/clickhouse-cpp/clickhouse/columns/bool.h
 - ext/clickhouse_native/vendor/clickhouse-cpp/clickhouse/columns/column.cpp
 - ext/clickhouse_native/vendor/clickhouse-cpp/clickhouse/columns/column.h
 - ext/clickhouse_native/vendor/clickhouse-cpp/clickhouse/columns/date.cpp
@@ -96,6 +108,8 @@ files:
 - ext/clickhouse_native/vendor/clickhouse-cpp/clickhouse/columns/ip6.h
 - ext/clickhouse_native/vendor/clickhouse-cpp/clickhouse/columns/itemview.cpp
 - ext/clickhouse_native/vendor/clickhouse-cpp/clickhouse/columns/itemview.h
+- ext/clickhouse_native/vendor/clickhouse-cpp/clickhouse/columns/json.cpp
+- ext/clickhouse_native/vendor/clickhouse-cpp/clickhouse/columns/json.h
 - ext/clickhouse_native/vendor/clickhouse-cpp/clickhouse/columns/lowcardinality.cpp
 - ext/clickhouse_native/vendor/clickhouse-cpp/clickhouse/columns/lowcardinality.h
 - ext/clickhouse_native/vendor/clickhouse-cpp/clickhouse/columns/lowcardinalityadaptor.h
@@ -310,7 +324,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 4.0.10
+rubygems_version: 4.0.16
 specification_version: 4
 summary: ClickHouse Ruby driver over the native TCP protocol
 test_files: []