iceberg 0.10.2-arm64-darwin → 0.11.0-arm64-darwin

data/README.md

@@ -2,6 +2,8 @@
 
 [Apache Iceberg](https://iceberg.apache.org/) for Ruby
 
+:duck: Also check out [SeaDuck](https://github.com/ankane/seaduck)
+
 [![Build Status](https://github.com/ankane/iceberg-ruby/actions/workflows/build.yml/badge.svg)](https://github.com/ankane/iceberg-ruby/actions)
 
 ## Installation
@@ -35,7 +37,7 @@ catalog.create_table("main.events") do |t|
 end
 ```
 
-Or
+Or with [Polars](https://github.com/ankane/ruby-polars)
 
 ```ruby
 df = Polars::DataFrame.new({"id" => [1, 2], "value" => [3.0, 4.0]})
@@ -82,9 +84,7 @@ Iceberg::MemoryCatalog.new(
 )
 ```
 
-## Reference
-
-### Namespaces
+## Namespaces
 
 List namespaces
 
@@ -122,7 +122,7 @@ Drop a namespace
 catalog.drop_namespace("main")
 ```
 
-### Tables
+## Tables
 
 List tables
 
@@ -169,7 +169,7 @@ Drop a table
 catalog.drop_table("main.events")
 ```
 
-### Static Tables
+## Static Tables
 
 Load a static table
 

data/lib/iceberg/catalog.rb

@@ -7,6 +7,7 @@ module Iceberg
     def create_namespace(namespace, properties: {}, if_not_exists: nil)
       @catalog.create_namespace(namespace, properties)
     rescue Error => e
+      # ideally all catalogs would use NamespaceAlreadyExistsError
       if !if_not_exists || (e.message != "Cannot create namespace" && !e.message.include?("already exists"))
         raise e
       end
@@ -28,7 +29,8 @@ module Iceberg
     def drop_namespace(namespace, if_exists: nil)
       @catalog.drop_namespace(namespace)
     rescue Error => e
-      if !if_exists || (e.message != "Tried to drop a namespace that does not exist" && !e.message.include?("No such namespace"))
+      # ideally all catalogs would use NamespaceNotFoundError
+      if !if_exists || (e.message != "Tried to drop a namespace that does not exist" && !e.message.include?("No such namespace") && !e.message.include?("The specified namespace does not exist") && !e.message.include?("not found"))
         raise e
       end
       nil
@@ -47,9 +49,9 @@ module Iceberg
         table_definition = TableDefinition.new
         yield table_definition
         schema = Schema.new(table_definition.fields)
-      elsif schema.is_a?(Hash)
+      elsif schema.is_a?(Hash) || (defined?(Polars::Schema) && schema.is_a?(Polars::Schema))
         fields =
-          schema.map.with_index do |(k, v), i|
+          schema.to_h.map.with_index do |(k, v), i|
             {
               id: i + 1,
               name: k.is_a?(Symbol) ? k.to_s : k,
@@ -72,7 +74,8 @@ module Iceberg
     def drop_table(table_name, if_exists: nil)
       @catalog.drop_table(table_name)
     rescue Error => e
-      if !if_exists || (e.message != "Tried to drop a table that does not exist" && !e.message.include?("No such table"))
+      # ideally all catalogs would use TableNotFoundError
+      if !if_exists || (e.message != "Tried to drop a table that does not exist" && !e.message.include?("No such table") && !e.message.include?("The specified table does not exist") && !e.message.include?("not found"))
         raise e
       end
       nil
@@ -92,11 +95,11 @@ module Iceberg
       @catalog.register_table(table_name, metadata_location)
     end
 
-    def query(sql)
+    def sql(sql)
       # requires datafusion feature
-      raise Todo unless @catalog.respond_to?(:query)
+      raise Todo unless @catalog.respond_to?(:sql)
 
-      @catalog.query(sql)
+      @catalog.sql(sql)
     end
 
     # hide internal state

data/lib/iceberg/glue_catalog.rb

@@ -2,9 +2,6 @@ module Iceberg
   class GlueCatalog < Catalog
     # warehouse is URI of S3 storage bucket
     def initialize(warehouse:)
-      # requires glue feature
-      raise Error, "Feature not enabled" unless RbCatalog.respond_to?(:new_glue)
-
       @catalog = RbCatalog.new_glue(warehouse)
     end
   end

data/lib/iceberg/s3_tables_catalog.rb

@@ -0,0 +1,7 @@
+module Iceberg
+  class S3TablesCatalog < Catalog
+    def initialize(arn:)
+      @catalog = RbCatalog.new_s3tables(arn)
+    end
+  end
+end

data/lib/iceberg/table.rb

@@ -83,56 +83,18 @@ module Iceberg
       @table.properties
     end
 
+    def scan(snapshot_id: nil)
+      TableScan.new(@table.scan(snapshot_id), self)
+    end
+
     def to_polars(snapshot_id: nil, storage_options: nil)
       require "polars-df"
 
-      files = @table.scan(snapshot_id).plan_files
-      if files.empty?
-        # TODO improve
-        schema =
-          # TODO use schema from snapshot_id
-          current_schema.fields.to_h do |field|
-            dtype =
-              case field[:type]
-              when "int"
-                Polars::Int32
-              when "long"
-                Polars::Int64
-              when "double"
-                Polars::Float64
-              when "string"
-                Polars::String
-              when "timestamp"
-                Polars::Datetime
-              else
-                raise Todo
-              end
-
-            [field[:name], dtype]
-          end
-
-        Polars::LazyFrame.new(schema: schema)
-      else
-        sources = files.map { |v| v[:data_file_path] }
-
-        deletion_files = [
-          "iceberg-position-delete",
-          files.map.with_index
-            .select { |v, i| v[:deletes].any? }
-            .to_h { |v, i| [i, v[:deletes].map { |d| d[:file_path] }] }
-        ]
-
-        Polars.scan_parquet(
-          sources,
-          storage_options: storage_options,
-          # TODO
-          # cast_options: Polars::ScanCastOptions._default_iceberg,
-          # allow_missing_columns: true,
-          # extra_columns: "ignore",
-          # _column_mapping: column_mapping,
-          _deletion_files: deletion_files
-        )
+      if Gem::Version.new(Polars::VERSION) < Gem::Version.new("0.23")
+        raise "Requires polars-df >= 0.23"
       end
+
+      Polars.scan_iceberg(self, snapshot_id:, storage_options:)
     end
 
     def append(df)

data/lib/iceberg/table_scan.rb

@@ -0,0 +1,18 @@
+module Iceberg
+  class TableScan
+    attr_reader :table
+
+    def initialize(scan, table)
+      @scan = scan
+      @table = table
+    end
+
+    def plan_files
+      @scan.plan_files
+    end
+
+    def snapshot
+      @scan.snapshot
+    end
+  end
+end

data/lib/iceberg/version.rb

@@ -1,3 +1,3 @@
 module Iceberg
-  VERSION = "0.10.2"
+  VERSION = "0.11.0"
 end

data/lib/iceberg.rb

@@ -9,6 +9,7 @@ end
 require_relative "iceberg/catalog"
 require_relative "iceberg/schema"
 require_relative "iceberg/table"
+require_relative "iceberg/table_scan"
 require_relative "iceberg/static_table"
 require_relative "iceberg/table_definition"
 require_relative "iceberg/version"
@@ -17,6 +18,7 @@ require_relative "iceberg/version"
 require_relative "iceberg/glue_catalog"
 require_relative "iceberg/memory_catalog"
 require_relative "iceberg/rest_catalog"
+require_relative "iceberg/s3_tables_catalog"
 require_relative "iceberg/sql_catalog"
 
 module Iceberg

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: iceberg
 version: !ruby/object:Gem::Version
-  version: 0.10.2
+  version: 0.11.0
 platform: arm64-darwin
 authors:
 - Andrew Kane
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2025-08-28 00:00:00.000000000 Z
+date: 2026-01-22 00:00:00.000000000 Z
 dependencies: []
 description: 
 email: andrew@ankane.org
@@ -24,18 +24,20 @@ files:
 - NOTICE.txt
 - README.md
 - lib/iceberg.rb
-- lib/iceberg/3.2/iceberg.bundle
 - lib/iceberg/3.3/iceberg.bundle
 - lib/iceberg/3.4/iceberg.bundle
+- lib/iceberg/4.0/iceberg.bundle
 - lib/iceberg/catalog.rb
 - lib/iceberg/glue_catalog.rb
 - lib/iceberg/memory_catalog.rb
 - lib/iceberg/rest_catalog.rb
+- lib/iceberg/s3_tables_catalog.rb
 - lib/iceberg/schema.rb
 - lib/iceberg/sql_catalog.rb
 - lib/iceberg/static_table.rb
 - lib/iceberg/table.rb
 - lib/iceberg/table_definition.rb
+- lib/iceberg/table_scan.rb
 - lib/iceberg/version.rb
 homepage: https://github.com/ankane/iceberg-ruby
 licenses:
@@ -49,10 +51,10 @@ required_ruby_version: !ruby/object:Gem::Requirement
   requirements:
   - - ">="
     - !ruby/object:Gem::Version
-      version: '3.2'
+      version: '3.3'
   - - "<"
     - !ruby/object:Gem::Version
-      version: 3.5.dev
+      version: 4.1.dev
 required_rubygems_version: !ruby/object:Gem::Requirement
   requirements:
   - - ">="