clickhouse-activerecord 0.5.0 → 0.5.7

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 1e17583f619debbb02952aa4299fedebbdd0de5ff6ae5c4a604071e42381c33a
-  data.tar.gz: d9d642a756c2ff7791081f1298917d96758dd5de4f07781dc7c9e587100536b4
+  metadata.gz: 79b43575e5eef93daa11e1a2aca2984f696cbef59021f55d1470b2dc0a5d1e3d
+  data.tar.gz: b79cccb85d8a07fa4dc533d7b7e9042eeaa2afead26214c075b3131e1843394b
 SHA512:
-  metadata.gz: cca1c4ce673a6622855a860e51b85952b0b09e3cbaaafe6b817a5c95ff1bb90d33da4974908acecaed6a658f7574fbaf9f71c1497d1c527006eb36bc3aaefab3
-  data.tar.gz: 344cc941b595fe40179c8974dfc371dd3732cbf8994999d75976506582d23abbda39340e7b2232cc7b7c25f1c6271d3f26325c409cfd7aa21f1acfc017fb8d32
+  metadata.gz: 840b01cc1d5b88eee5e031fada23cfd35f7a8d948a35a767b2f5baf1596af665f28c56778e699960511004367b16ee3130be1f5bd2ee713a9a0d5053a9b58a37
+  data.tar.gz: 4e613a19c51b362a05634da06d9528809ed09f01ebec4eac9984656156e9b877bdc4ce321e24bea63ae4e49d85609fa9dda781db60159e60b6d6ee06e8a99a3e

data/CHANGELOG.md

@@ -1,3 +1,15 @@
+### Version 0.5.6 (Oct 25, 2021)
+ 
+ * Added auto creating service distributed tables and additional options for creating view [@ygreeek](https://github.com/ygreeek)
+ * Added default user agent
+
+### Version 0.5.3 (Sep 22, 2021)
+ 
+ * Fix replica cluster for a new syntax MergeTree
+ * Fix support rails 5.2 on alter table
+ * Support array type of column
+ * Support Rails 6.1.0 [@bdevel](https://github.com/bdevel)
+
 ### Version 0.4.10 (Mar 10, 2021)
 
  * Support ClickHouse 20.9+

data/README.md

@@ -196,6 +196,7 @@ false`. The default integer is `UInt32`
 | UInt32 | 0 to 4,294,967,295 | 3,4 |
 | UInt64 | 0 to 18446744073709551615 | 5,6,7,8 |
 | UInt256 | 0 to ... | 8+ |
+| Array | ... | ... |
 
 Example:
 

data/core_extensions/active_record/migration/command_recorder.rb

@@ -0,0 +1,27 @@
+module CoreExtensions
+  module ActiveRecord
+    module Migration
+      module CommandRecorder
+        def create_table_with_distributed(*args, &block)
+          record(:create_table_with_distributed, args, &block)
+        end
+
+        def create_view(*args, &block)
+          record(:create_view, args, &block)
+        end
+
+        private
+
+        def invert_create_table_with_distributed(args)
+          table_name, options = args
+          [:drop_table_with_distributed, table_name, options]
+        end
+
+        def invert_create_view(args)
+          view_name, options = args
+          [:drop_table, view_name, options]
+        end
+      end
+    end
+  end
+end

data/lib/active_record/connection_adapters/clickhouse/oid/array.rb

@@ -0,0 +1,30 @@
+# frozen_string_literal: true
+
+module ActiveRecord
+  module ConnectionAdapters
+    module Clickhouse
+      module OID # :nodoc:
+        class Array < Type::Value # :nodoc:
+
+          def initialize(sql_type)
+            @subtype = case sql_type
+                       when /U?Int\d+/
+                         :integer
+                       when /DateTime/
+                         :datetime
+                       when /Date/
+                         :date
+                       else
+                         :string
+            end
+          end
+
+          def type
+            @subtype
+          end
+
+        end
+      end
+    end
+  end
+end

data/lib/active_record/connection_adapters/clickhouse/schema_creation.rb

@@ -22,6 +22,9 @@ module ActiveRecord
           if options[:null] || options[:null].nil?
             sql.gsub!(/\s+(.*)/, ' Nullable(\1)')
           end
+          if options[:array]
+            sql.gsub!(/\s+(.*)/, ' Array(\1)')
+          end
           sql.gsub!(/(\sString)\(\d+\)/, '\1')
           sql << " DEFAULT #{quote_default_expression(options[:default], options[:column])}" if options_include_default?(options)
           sql
@@ -33,7 +36,7 @@ module ActiveRecord
             # rails 6.1
             opts ||= options.options
           end
-          
+
           if opts.present?
             create_sql << " ENGINE = #{opts}"
           else
@@ -43,17 +46,48 @@ module ActiveRecord
           create_sql
         end
 
+        def add_as_clause!(create_sql, options)
+          return unless options.as
+
+          assign_database_to_subquery!(options.as) if options.view
+          create_sql << " AS #{to_sql(options.as)}"
+        end
+
+        def assign_database_to_subquery!(subquery)
+          # If you do not specify a database explicitly, ClickHouse will use the "default" database.
+          return unless subquery
+
+          match = subquery.match(/(?<=from)[^.\w]+(?<database>\w+(?=\.))?(?<table_name>[.\w]+)/i)
+          return unless match
+          return if match[:database]
+
+          subquery[match.begin(:table_name)...match.end(:table_name)] =
+            "#{current_database}.#{match[:table_name].sub('.', '')}"
+        end
+
+        def add_to_clause!(create_sql, options)
+          # If you do not specify a database explicitly, ClickHouse will use the "default" database.
+          return unless options.to
+
+          match = options.to.match(/(?<database>.+(?=\.))?(?<table_name>.+)/i)
+          return unless match
+          return if match[:database]
+
+          create_sql << "TO #{current_database}.#{options.to.sub('.', '')} "
+        end
+
         def visit_TableDefinition(o)
           create_sql = +"CREATE#{table_modifier_in_create(o)} #{o.view ? "VIEW" : "TABLE"} "
           create_sql << "IF NOT EXISTS " if o.if_not_exists
           create_sql << "#{quote_table_name(o.name)} "
+          add_to_clause!(create_sql, o) if o.materialized
 
           statements = o.columns.map { |c| accept c }
           statements << accept(o.primary_keys) if o.primary_keys
           create_sql << "(#{statements.join(', ')})" if statements.present?
-          # Attach options for only table or materialized view
-          add_table_options!(create_sql, o)  if !o.view || o.view && o.materialized
-          create_sql << " AS #{to_sql(o.as)}" if o.as
+          # Attach options for only table or materialized view without TO section
+          add_table_options!(create_sql, o) if !o.view || o.view && o.materialized && !o.to
+          add_as_clause!(create_sql, o)
           create_sql
         end
 
@@ -81,6 +115,9 @@ module ActiveRecord
           change_column_sql
         end
 
+        def current_database
+          ActiveRecord::Base.connection_db_config.database
+        end
       end
     end
   end

data/lib/active_record/connection_adapters/clickhouse/schema_definitions.rb

@@ -5,7 +5,7 @@ module ActiveRecord
     module Clickhouse
       class TableDefinition < ActiveRecord::ConnectionAdapters::TableDefinition
 
-        attr_reader :view, :materialized, :if_not_exists
+        attr_reader :view, :materialized, :if_not_exists, :to
 
         def initialize(
             conn,
@@ -17,6 +17,7 @@ module ActiveRecord
             comment: nil,
             view: false,
             materialized: false,
+            to: nil,
             **
           )
           @conn = conn
@@ -32,6 +33,7 @@ module ActiveRecord
           @comment = comment
           @view = view || materialized
           @materialized = materialized
+          @to = to
         end
 
         def integer(*args, **options)
@@ -58,7 +60,7 @@ module ActiveRecord
               kind = :int256     if options[:limit] > 16
             end
           end
-          args.each { |name| column(name, kind, options.except(:limit, :unsigned)) }
+          args.each { |name| column(name, kind, **options.except(:limit, :unsigned)) }
         end
       end
     end

data/lib/active_record/connection_adapters/clickhouse/schema_statements.rb

@@ -1,5 +1,7 @@
 # frozen_string_literal: true
 
+require 'clickhouse-activerecord/version'
+
 module ActiveRecord
   module ConnectionAdapters
     module Clickhouse
@@ -23,6 +25,11 @@ module ActiveRecord
           raise ActiveRecord::ActiveRecordError, "Response: #{e.message}"
         end
 
+        def exec_insert_all(sql, name)
+          do_execute(sql, name, format: nil)
+          true
+        end
+
         def exec_update(_sql, _name = nil, _binds = [])
           raise ActiveRecord::ActiveRecordError, 'Clickhouse update is not supported'
         end
@@ -53,7 +60,7 @@ module ActiveRecord
 
         def do_system_execute(sql, name = nil)
           log_with_debug(sql, "#{adapter_name} #{name}") do
-            res = @connection.post("/?#{@config.to_param}", "#{sql} FORMAT JSONCompact")
+            res = @connection.post("/?#{@config.to_param}", "#{sql} FORMAT JSONCompact", 'User-Agent' => "Clickhouse ActiveRecord #{ClickhouseActiverecord::VERSION}")
 
             process_response(res)
           end
@@ -63,7 +70,7 @@ module ActiveRecord
           log(sql, "#{adapter_name} #{name}") do
             formatted_sql = apply_format(sql, format)
             request_params = @config || {}
-            res = @connection.post("/?#{request_params.merge(settings).to_param}", formatted_sql)
+            res = @connection.post("/?#{request_params.merge(settings).to_param}", formatted_sql, 'User-Agent' => "Clickhouse ActiveRecord #{ClickhouseActiverecord::VERSION}")
 
             process_response(res)
           end
@@ -116,7 +123,7 @@ module ActiveRecord
           Clickhouse::SchemaCreation.new(self)
         end
 
-        def create_table_definition(table_name, options)
+        def create_table_definition(table_name, **options)
           Clickhouse::TableDefinition.new(self, table_name, **options)
         end
 

data/lib/active_record/connection_adapters/clickhouse_adapter.rb

@@ -3,6 +3,7 @@
 require 'clickhouse-activerecord/arel/visitors/to_sql'
 require 'clickhouse-activerecord/arel/table'
 require 'clickhouse-activerecord/migration'
+require 'active_record/connection_adapters/clickhouse/oid/array'
 require 'active_record/connection_adapters/clickhouse/oid/date'
 require 'active_record/connection_adapters/clickhouse/oid/date_time'
 require 'active_record/connection_adapters/clickhouse/oid/big_integer'
@@ -17,6 +18,7 @@ module ActiveRecord
       # Establishes a connection to the database that's used by all Active Record objects
       def clickhouse_connection(config)
         config = config.symbolize_keys
+
         if config[:connection]
           connection = {
             connection: config[:connection]
@@ -27,6 +29,9 @@ module ActiveRecord
             host: config[:host] || 'localhost',
             port: port,
             ssl: config[:ssl].present? ? config[:ssl] : port == 443,
+            sslca: config[:sslca],
+            read_timeout: config[:read_timeout],
+            write_timeout: config[:write_timeout],
           }
         end
 
@@ -190,12 +195,16 @@ module ActiveRecord
         register_class_with_limit m, %r(Int128), Type::Integer
         register_class_with_limit m, %r(Int256), Type::Integer
 
-        register_class_with_limit m, %r(Uint8), Type::UnsignedInteger
+        register_class_with_limit m, %r(UInt8), Type::UnsignedInteger
         register_class_with_limit m, %r(UInt16), Type::UnsignedInteger
         register_class_with_limit m, %r(UInt32), Type::UnsignedInteger
         register_class_with_limit m, %r(UInt64), Type::UnsignedInteger
         #register_class_with_limit m, %r(UInt128), Type::UnsignedInteger #not implemnted in clickhouse
         register_class_with_limit m, %r(UInt256), Type::UnsignedInteger
+        # register_class_with_limit m, %r(Array), Clickhouse::OID::Array
+        m.register_type(%r(Array)) do |sql_type|
+          Clickhouse::OID::Array.new(sql_type)
+        end
       end
 
       # Quoting time without microseconds
@@ -253,7 +262,7 @@ module ActiveRecord
       def create_view(table_name, **options)
         options.merge!(view: true)
         options = apply_replica(table_name, options)
-        td = create_table_definition(apply_cluster(table_name), options)
+        td = create_table_definition(apply_cluster(table_name), **options)
         yield td if block_given?
 
         if options[:force]
@@ -263,10 +272,10 @@ module ActiveRecord
         execute schema_creation.accept td
       end
 
-      def create_table(table_name, **options)
+      def create_table(table_name, **options, &block)
         options = apply_replica(table_name, options)
-        td = create_table_definition(apply_cluster(table_name), options)
-        yield td if block_given?
+        td = create_table_definition(apply_cluster(table_name), **options)
+        block.call td if block_given?
 
         if options[:force]
           drop_table(table_name, options.merge(if_exists: true))
@@ -275,6 +284,19 @@ module ActiveRecord
         execute schema_creation.accept td
       end
 
+      def create_table_with_distributed(table_name, **options, &block)
+        sharding_key = options.delete(:sharding_key) || 'rand()'
+        create_table("#{table_name}_distributed", **options, &block)
+        raise 'Set a cluster' unless cluster
+
+        distributed_options = "Distributed(#{cluster},#{@config[:database]},#{table_name}_distributed,#{sharding_key})"
+        create_table(table_name, **options.merge(options: distributed_options), &block)
+      end
+
+      def drop_table_with_distributed(table_name, **options)
+        ["#{table_name}_distributed", table_name].each { |name| drop_table(name, **options) }
+      end
+
       # Drops a ClickHouse database.
       def drop_database(name) #:nodoc:
         sql = apply_cluster "DROP DATABASE IF EXISTS #{quote_table_name(name)}"
@@ -315,14 +337,41 @@ module ActiveRecord
         @full_config[:replica_name]
       end
 
+      def use_default_replicated_merge_tree_params?
+        database_engine_atomic? && @full_config[:use_default_replicated_merge_tree_params]
+      end
+
+      def use_replica?
+        (replica || use_default_replicated_merge_tree_params?) && cluster
+      end
+
       def replica_path(table)
         "/clickhouse/tables/#{cluster}/#{@config[:database]}.#{table}"
       end
 
+      def database_engine_atomic?
+        current_database_engine = "select engine from system.databases where name = '#{@config[:database]}'"
+        res = ActiveRecord::Base.connection.select_one(current_database_engine)
+        res['engine'] == 'Atomic' if res
+      end
+
       def apply_cluster(sql)
         cluster ? "#{sql} ON CLUSTER #{cluster}" : sql
       end
 
+      def supports_insert_on_duplicate_skip?
+        true
+      end
+
+      def supports_insert_on_duplicate_update?
+        true
+      end
+
+      def build_insert_sql(insert) # :nodoc:
+        sql = +"INSERT #{insert.into} #{insert.values_list}"
+        sql
+      end
+
       protected
 
       def last_inserted_id(result)
@@ -339,17 +388,35 @@ module ActiveRecord
 
       def connect
         @connection = @connection_parameters[:connection] || Net::HTTP.start(@connection_parameters[:host], @connection_parameters[:port], use_ssl: @connection_parameters[:ssl], verify_mode: OpenSSL::SSL::VERIFY_NONE)
+
+        @connection.ca_file = @connection_parameters[:ca_file] if @connection_parameters[:ca_file]
+        @connection.read_timeout = @connection_parameters[:read_timeout] if @connection_parameters[:read_timeout]
+        @connection.write_timeout = @connection_parameters[:write_timeout] if @connection_parameters[:write_timeout]
+
+        @connection
       end
 
       def apply_replica(table, options)
-        if replica && cluster && options[:options]
-          match = options[:options].match(/^(.*?MergeTree)\(([^\)]*)\)(.*?)$/)
-          if match
-            options[:options] = "Replicated#{match[1]}(#{([replica_path(table), replica].map{|v| "'#{v}'"} + [match[2].presence]).compact.join(', ')})#{match[3]}"
+        if use_replica? && options[:options]
+          if options[:options].match(/^Replicated/)
+            raise 'Do not try create Replicated table. It will be configured based on the *MergeTree engine.'
           end
+
+          options[:options] = configure_replica(table, options[:options])
         end
         options
       end
+
+      def configure_replica(table, options)
+        match = options.match(/^(.*?MergeTree)(?:\(([^\)]*)\))?((?:.|\n)*)/)
+        return options unless match
+
+        if replica
+          engine_params = ([replica_path(table), replica].map { |v| "'#{v}'" } + [match[2].presence]).compact.join(', ')
+        end
+
+        "Replicated#{match[1]}(#{engine_params})#{match[3]}"
+      end
     end
   end
 end

data/lib/clickhouse-activerecord/migration.rb

@@ -6,14 +6,23 @@ module ClickhouseActiverecord
     class << self
 
       def create_table
-        unless table_exists?
-          version_options = connection.internal_string_options_for_primary_key
-
-          connection.create_table(table_name, id: false, options: 'ReplacingMergeTree(ver) PARTITION BY version ORDER BY (version)', if_not_exists: true) do |t|
-            t.string :version, **version_options
-            t.column :active, 'Int8', null: false, default: '1'
-            t.datetime :ver, null: false, default: -> { 'now()' }
-          end
+        return if table_exists?
+
+        version_options = connection.internal_string_options_for_primary_key
+        table_options = {
+          id: false, options: 'ReplacingMergeTree(ver) PARTITION BY version ORDER BY (version)', if_not_exists: true
+        }
+        if connection.instance_variable_get(:@full_config)[:distributed_service_tables]
+          table_options.merge!(sharding_key: 'cityHash64(version)')
+          table_creation_method = 'create_table_with_distributed'
+        else
+          table_creation_method = 'create_table'
+        end
+
+        connection.public_send(table_creation_method, table_name, **table_options) do |t|
+          t.string :version, **version_options
+          t.column :active, 'Int8', null: false, default: '1'
+          t.datetime :ver, null: false, default: -> { 'now()' }
         end
       end
 
@@ -26,14 +35,25 @@ module ClickhouseActiverecord
   class InternalMetadata < ::ActiveRecord::InternalMetadata
     class << self
       def create_table
-        unless table_exists?
-          key_options = connection.internal_string_options_for_primary_key
-
-          connection.create_table(table_name, id: false, options: connection.adapter_name.downcase == 'clickhouse' ? 'MergeTree() PARTITION BY toDate(created_at) ORDER BY (created_at)' : '', if_not_exists: true) do |t|
-            t.string :key, key_options
-            t.string :value
-            t.timestamps
-          end
+        return if table_exists?
+
+        key_options = connection.internal_string_options_for_primary_key
+        table_options = {
+          id: false,
+          options: connection.adapter_name.downcase == 'clickhouse' ? 'MergeTree() PARTITION BY toDate(created_at) ORDER BY (created_at)' : '',
+          if_not_exists: true
+        }
+        if connection.instance_variable_get(:@full_config).try(:[], :distributed_service_tables)
+          table_options.merge!(sharding_key: 'cityHash64(created_at)')
+          table_creation_method = 'create_table_with_distributed'
+        else
+          table_creation_method = 'create_table'
+        end
+
+        connection.public_send(table_creation_method, table_name, **table_options) do |t|
+          t.string :key, **key_options
+          t.string :value
+          t.timestamps
         end
       end
     end
@@ -47,6 +67,16 @@ module ClickhouseActiverecord
       @schema_migration = schema_migration
     end
 
+    def up(target_version = nil)
+      selected_migrations = if block_given?
+        migrations.select { |m| yield m }
+      else
+        migrations
+      end
+
+      ClickhouseActiverecord::Migrator.new(:up, selected_migrations, schema_migration, target_version).migrate
+    end
+
     def down(target_version = nil)
       selected_migrations = if block_given?
         migrations.select { |m| yield m }

data/lib/clickhouse-activerecord/schema_dumper.rb

@@ -144,9 +144,14 @@ HEADER
       (column.sql_type =~ /(Nullable)?\(?UInt\d+\)?/).nil? ? false : nil
     end
 
+    def schema_array(column)
+      (column.sql_type =~ /Array?\(/).nil? ? nil : true
+    end
+
     def prepare_column_options(column)
       spec = {}
       spec[:unsigned] = schema_unsigned(column)
+      spec[:array] = schema_array(column)
       spec.merge(super).compact
     end
   end

data/lib/clickhouse-activerecord/tasks.rb

@@ -43,7 +43,15 @@ module ClickhouseActiverecord
     end
 
     def structure_load(*args)
-      File.read(args.first).split(";\n\n").each { |sql| connection.execute(sql) }
+      File.read(args.first).split(";\n\n").each do |sql|
+        if sql.gsub(/[a-z]/i, '').blank?
+          next
+        elsif sql =~ /^INSERT INTO/
+          connection.do_execute(sql, nil, format: nil)
+        else
+          connection.execute(sql)
+        end
+      end
     end
 
     def migrate

data/lib/clickhouse-activerecord/version.rb

@@ -1,3 +1,3 @@
 module ClickhouseActiverecord
-  VERSION = '0.5.0'
+  VERSION = '0.5.7'
 end

data/lib/clickhouse-activerecord.rb

@@ -2,6 +2,9 @@
 
 require 'active_record/connection_adapters/clickhouse_adapter'
 
+require_relative '../core_extensions/active_record/migration/command_recorder'
+ActiveRecord::Migration::CommandRecorder.include CoreExtensions::ActiveRecord::Migration::CommandRecorder
+
 if defined?(Rails::Railtie)
   require 'clickhouse-activerecord/railtie'
   require 'clickhouse-activerecord/schema'

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: clickhouse-activerecord
 version: !ruby/object:Gem::Version
-  version: 0.5.0
+  version: 0.5.7
 platform: ruby
 authors:
 - Sergey Odintsov
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2021-06-10 00:00:00.000000000 Z
+date: 2021-10-25 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler
@@ -112,6 +112,8 @@ files:
 - bin/console
 - bin/setup
 - clickhouse-activerecord.gemspec
+- core_extensions/active_record/migration/command_recorder.rb
+- lib/active_record/connection_adapters/clickhouse/oid/array.rb
 - lib/active_record/connection_adapters/clickhouse/oid/big_integer.rb
 - lib/active_record/connection_adapters/clickhouse/oid/date.rb
 - lib/active_record/connection_adapters/clickhouse/oid/date_time.rb