egis 1.5.0 → 2.0.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 924727f52ac809df7ab4457a6a0c05469cd6c128adaff339018e964dc9ff4ea0
-  data.tar.gz: c8ba9d090641c899e832b73b062314bd76fa4fb39041fb175d7e3a717e943856
+  metadata.gz: ca7464c31cc32115edb77470e2d5f2abe99ab9f4a64eef2ae0820ed8d5b982b7
+  data.tar.gz: 8002e611e1c59635365a3ca2ff54191e1a9671f4b5498c940438c85b7ea45d28
 SHA512:
-  metadata.gz: fa429f550a68475a75f9dd9c0610b0d41c9a8017f0e46c0f9e467cfaeddfb7d6587d6e7df5b4440bbd75e4d4d9a12d5966f40e4ce72cb8abf70e5dc2897f91d8
-  data.tar.gz: 3beac29ce82c267bfc2621e730a8b631f76a94f7426ac7d7901b31d204cda3320d1ab934e9c3e156ee52e5fca315728412ffd5a007bf4623dd2ac077d85399ba
+  metadata.gz: 5971ac98ab6bfdcbcac23a4b9e0e2b8b31b14cb06987ea6f3891fc0c1787087ebfca85eae026c8cb1c2952f11e4659cf36827492b8d9fdf7d43d3a6943d3bd7f
+  data.tar.gz: 98d379a3ddb1b3796b08e9c5a4cc584826324fa03e24d64ddb604405ec57a6f0d75a3d2138cf9123f154e275b5202c592814c4cea6a00049d75eda6ec5d2de8b

data/lib/egis/database.rb

@@ -28,8 +28,7 @@ module Egis
     # @param [String] table_name
     # @param [Egis::TableSchema] table_schema
     # @param [String] table_location S3 URL with table location (e.g. `s3://s3_bucket/table/location/`)
-    # @param [:tsv, :csv, :orc, {serde: 'SerdeClass', serde_properties: {property: value}}] format Table format
-    # (defaults to :tsv)
+    # @param [:tsv, :csv, :orc, :orc_index_access, :json, String] format Table Format (defaults to :tsv)
     # @return [Egis::Table]
 
     def table(table_name, table_schema, table_location, **options)

data/lib/egis/table.rb

@@ -142,7 +142,6 @@ module Egis
 
     ##
     # @return Table data format
-
     def format
       options.fetch(:format)
     end

data/lib/egis/table_ddl_generator.rb

@@ -9,7 +9,7 @@ module Egis
           #{column_definition_sql(table.schema.columns)}
         )
         #{partition_statement(table.schema)}
-        #{row_format_statement(table.format)}
+        #{format_statement(table.format)}
         LOCATION '#{table.location}';
       SQL
     end
@@ -34,37 +34,31 @@ module Egis
       columns.map { |column| "`#{column.name}` #{column.type}" }.join(",\n")
     end
 
-    def serde?(format)
-      format.is_a?(Hash) && format.key?(:serde)
-    end
-
-    def row_format_statement(format)
-      return serde_row_format_statement(format) if serde?(format)
-
-      delimited_row_format_statement(format)
-    end
+    def format_statement(format)
+      return format if format.is_a?(String)
 
-    def serde_row_format_statement(format)
-      row_format = "ROW FORMAT SERDE '#{format[:serde]}'"
-      return row_format unless format.key?(:serde_properties)
-
-      serde_properties = format[:serde_properties].map { |property, value| "'#{property}' = '#{value}'" }
-      <<-SQL
-        #{row_format}
-        WITH SERDEPROPERTIES (
-          #{serde_properties.join(",\n")}
-        )
-      SQL
+      format_preset(format)
     end
 
-    def delimited_row_format_statement(format)
+    def format_preset(format) # rubocop:disable Metrics/MethodLength
       case format
       when :csv
         "ROW FORMAT DELIMITED FIELDS TERMINATED BY ','"
       when :tsv
         "ROW FORMAT DELIMITED FIELDS TERMINATED BY '\\t'"
       when :orc
+        <<~SQL
+          ROW FORMAT SERDE 'org.apache.hadoop.hive.ql.io.orc.OrcSerde'
+          WITH SERDEPROPERTIES (
+            'orc.column.index.access' = 'false'
+          )
+          STORED AS INPUTFORMAT 'org.apache.hadoop.hive.ql.io.orc.OrcInputFormat'
+          OUTPUTFORMAT 'org.apache.hadoop.hive.ql.io.orc.OrcOutputFormat'
+        SQL
+      when :orc_index_access
         'STORED AS ORC'
+      when :json
+        "ROW FORMAT SERDE 'org.apache.hive.hcatalog.data.JsonSerDe'"
       else
         raise Errors::UnsupportedTableFormat, format.to_s
       end

data/lib/egis/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Egis
-  VERSION = '1.5.0'
+  VERSION = '2.0.0'
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: egis
 version: !ruby/object:Gem::Version
-  version: 1.5.0
+  version: 2.0.0
 platform: ruby
 authors:
 - Agnieszka Czereba
@@ -9,7 +9,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2021-05-14 00:00:00.000000000 Z
+date: 2021-12-16 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: aws-sdk-athena