parquet 0.5.1-arm64-darwin → 0.5.3-arm64-darwin

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 4699979e4f19d30fc231e9b5a786696033a10695e2b8eed42cedbfd166b3242f
-  data.tar.gz: 49a7d0b6aa605c709ca36b7cf3abe5bbcaff7d2c2a920b74afe9400dd6bf5259
+  metadata.gz: b7eb67be4612a8c8d5e197ba5b054264e97a428409d6c090c743d70b608b3308
+  data.tar.gz: 8c07288afe0c2211306c7785557fd92429fd954f9ef1861456bae610de10c2b5
 SHA512:
-  metadata.gz: 5640d3e1d37205cb442b4193d9639edc9ceabb806a0a07fcc63c74b8fc1636ae0f17387d90707aafb8cd2e0533df9e9378ef3d923d7436ddeac721fe89c0220e
-  data.tar.gz: c679658ca9341ed0b74a29725c47f65e9dbbf2fe5b1893380723fe396479987407e295fa2c41535c4dd2550d419497ec335a84f572a9d7873411db6ace33be19
+  metadata.gz: eb9693596f72f2659dae5caffe35c13a781b3231589875ddaba9a22fa16e1d6a3901bf8ced9e84ec1ab2991b759b48643d16da2af5a7042c218f89482727d907
+  data.tar.gz: b7695b0dfea66a3535961caf69f5f9f7b36db31a54e56086b803d39d6bba7e19dc4a066bc3dd58dc7c28b3d312d12c30923555e675d2445b677981e3696a629a

data/lib/parquet/schema.rb

@@ -11,6 +11,9 @@ module Parquet
     #     field :id, :int64, nullable: false  # ID cannot be null
     #     field :name, :string  # Default nullable: true
     #
+    #     # Decimal field with precision and scale
+    #     field :price, :decimal, precision: 10, scale: 2
+    #
     #     # List with non-nullable items
     #     field :scores, :list, item: :float, item_nullable: false
     #
@@ -45,7 +48,7 @@ module Parquet
 
       # Define a field in the schema
       # @param name [String, Symbol] field name
-      # @param type [Symbol] data type (:int32, :int64, :string, :list, :map, :struct, etc)
+      # @param type [Symbol] data type (:int32, :int64, :string, :list, :map, :struct, :decimal, etc)
       # @param nullable [Boolean] whether the field can be null (default: true)
       # @param kwargs [Hash] additional options depending on type
       #
@@ -55,6 +58,7 @@ module Parquet
       #   - `key:, value:` if type == :map
       #   - `key_nullable:, value_nullable:` controls nullability of map keys/values (default: true)
       #   - `format:` if you want to store some format string
+      #   - `precision:, scale:` if type == :decimal (precision defaults to 18, scale to 2)
       #   - `nullable:` default to true if not specified
       def field(name, type, nullable: true, **kwargs, &block)
         field_hash = { name: name.to_s, type: type, nullable: !!nullable }
@@ -73,7 +77,15 @@ module Parquet
           raise ArgumentError, "list field `#{name}` requires `item:` type" unless item_type
           # Pass item_nullable if provided, otherwise use true as default
           item_nullable = kwargs[:item_nullable].nil? ? true : !!kwargs[:item_nullable]
-          field_hash[:item] = wrap_subtype(item_type, nullable: item_nullable, &block)
+          
+          # Pass precision and scale if type is decimal
+          if item_type == :decimal
+            precision = kwargs[:precision]
+            scale = kwargs[:scale]
+            field_hash[:item] = wrap_subtype(item_type, nullable: item_nullable, precision: precision, scale: scale, &block)
+          else
+            field_hash[:item] = wrap_subtype(item_type, nullable: item_nullable, &block)
+          end
         when :map
           # user must specify key:, value:
           key_type = kwargs[:key]
@@ -82,8 +94,41 @@ module Parquet
           # Pass key_nullable and value_nullable if provided, otherwise use true as default
           key_nullable = kwargs[:key_nullable].nil? ? true : !!kwargs[:key_nullable]
           value_nullable = kwargs[:value_nullable].nil? ? true : !!kwargs[:value_nullable]
+          
           field_hash[:key] = wrap_subtype(key_type, nullable: key_nullable)
-          field_hash[:value] = wrap_subtype(value_type, nullable: value_nullable, &block)
+          
+          # Pass precision and scale if value type is decimal
+          if value_type == :decimal
+            precision = kwargs[:precision]
+            scale = kwargs[:scale]
+            field_hash[:value] = wrap_subtype(value_type, nullable: value_nullable, precision: precision, scale: scale, &block)
+          else
+            field_hash[:value] = wrap_subtype(value_type, nullable: value_nullable, &block)
+          end
+        when :decimal
+          # Store precision and scale for decimal type according to rules:
+          # 1. When neither precision nor scale is provided, use maximum precision (38)
+          # 2. When only precision is provided, scale defaults to 0
+          # 3. When only scale is provided, use maximum precision (38)
+          # 4. When both are provided, use the provided values
+          
+          if kwargs[:precision].nil? && kwargs[:scale].nil?
+            # No precision or scale provided - use maximum precision
+            field_hash[:precision] = 38
+            field_hash[:scale] = 0
+          elsif kwargs[:precision] && kwargs[:scale].nil?
+            # Precision only - scale defaults to 0
+            field_hash[:precision] = kwargs[:precision]
+            field_hash[:scale] = 0
+          elsif kwargs[:precision].nil? && kwargs[:scale]
+            # Scale only - use maximum precision
+            field_hash[:precision] = 38
+            field_hash[:scale] = kwargs[:scale]
+          else
+            # Both provided
+            field_hash[:precision] = kwargs[:precision]
+            field_hash[:scale] = kwargs[:scale]
+          end
         else
           # primitive type: :int32, :int64, :string, etc.
           # do nothing else special
@@ -122,7 +167,7 @@ module Parquet
       # If user said: field "something", :list, item: :struct do ... end
       # we want to recursively parse that sub-struct from the block.
       # So wrap_subtype might be:
-      def wrap_subtype(t, nullable: true, &block)
+      def wrap_subtype(t, nullable: true, precision: nil, scale: nil, &block)
         if t == :struct
           sub_builder = SchemaBuilder.new
           sub_builder.instance_eval(&block) if block
@@ -144,6 +189,34 @@ module Parquet
           end
 
           { type: :list, nullable: nullable, name: "item", item: sub_builder.fields[0] }
+        elsif t == :decimal
+          # Handle decimal type with precision and scale
+          result = { type: t, nullable: nullable, name: "item" }
+          
+          # Follow the same rules as in field() method:
+          # 1. When neither precision nor scale is provided, use maximum precision (38)
+          # 2. When only precision is provided, scale defaults to 0
+          # 3. When only scale is provided, use maximum precision (38)
+          # 4. When both are provided, use the provided values
+          if precision.nil? && scale.nil?
+            # No precision or scale provided - use maximum precision
+            result[:precision] = 38
+            result[:scale] = 0
+          elsif precision && scale.nil?
+            # Precision only - scale defaults to 0
+            result[:precision] = precision
+            result[:scale] = 0
+          elsif precision.nil? && scale
+            # Scale only - use maximum precision
+            result[:precision] = 38
+            result[:scale] = scale
+          else
+            # Both provided
+            result[:precision] = precision
+            result[:scale] = scale
+          end
+          
+          result
         else
           # e.g. :int32 => { type: :int32, nullable: true }
           { type: t, nullable: nullable, name: "item" }

data/lib/parquet/version.rb

@@ -1,3 +1,3 @@
 module Parquet
-  VERSION = "0.5.1"
+  VERSION = "0.5.3"
 end

data/lib/parquet.rbi

@@ -1,6 +1,17 @@
 # typed: true
 
 module Parquet
+  # Returns metadata information about a Parquet file
+  #
+  # The returned hash contains information about:
+  # - Basic file metadata (num_rows, created_by)
+  # - Schema information (fields, types, etc.)
+  # - Row group details
+  # - Column chunk information (compression, encodings, statistics)
+  sig { params(path: String).returns(T::Hash[String, T.untyped]) }
+  def self.metadata(path)
+  end
+
   # Options:
   #   - `input`: String, File, or IO object containing parquet data
   #   - `result_type`: String specifying the output format

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: parquet
 version: !ruby/object:Gem::Version
-  version: 0.5.1
+  version: 0.5.3
 platform: arm64-darwin
 authors:
 - Nathan Jaremko
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2025-02-27 00:00:00.000000000 Z
+date: 2025-04-01 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rake-compiler