data_drain 0.1.4 → 0.1.6

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 9e6e7cad34a499305184a64e46fe48391360a16647defdd8c566f087a3c90fac
-  data.tar.gz: a8cb4dbd541b0f4b1e4dbb58319c5112e66f58b581bb2f7c6771c4b1668160c4
+  metadata.gz: eb6e8dad588e5512397bb9ed8f99b50bb32dcacf6eb493fbef2dfcf0cb4f8289
+  data.tar.gz: 241eb6c7e340b25d95757f0e663489eba65cb1b9293426c81148c45c78ea534f
 SHA512:
-  metadata.gz: 4d7ee0aa8f6fb83eb9accd969cdde65e71e8e4d0393c5e346554f78174587af5453d2e3f0cd1314a04432a618a682eff6985ac79e14c7e06d3e83ada409e48b3
-  data.tar.gz: aa75956d57b72d2a9ad04d3e05ebf7603b4bc58eaaf46552ac11940c4e18e31d3eab994412ad6581133bccba0a280b057925135c5b2c86a1e0067bdc12fb7319
+  metadata.gz: 04ca70eca218c7a9aadd292c4b8def9b4895c2b223a2d46f4297fdd4d078b1d8fd05153ce7410fda35f6783242dddfc1a25993c3b2b3db1aad81572da119cdf2
+  data.tar.gz: dd168e87460ac02365bbc7f1d9f80050c6a103de1f8f0820545933458cce7aa736c6edb4f9c5e8292f436253082ab43ba874182b14709aeb71430cece9a6f56b

data/CHANGELOG.md

@@ -1,5 +1,13 @@
 ## [Unreleased]
 
+## [0.1.6] - 2026-03-16
+
+- Se agrego el tem_directory para duckdb.
+
+## [0.1.5] - 2026-03-16
+
+- Se agrego el attach para duckdb.
+
 ## [0.1.4] - 2026-03-16
 
 - Corrección de error por comilla simple

data/lib/data_drain/configuration.rb

@@ -8,7 +8,7 @@ module DataDrain
     attr_accessor :storage_mode, :aws_region,
       :aws_access_key_id, :aws_secret_access_key,
       :db_host, :db_port, :db_user, :db_pass, :db_name,
-      :batch_size, :throttle_delay, :logger, :limit_ram
+      :batch_size, :throttle_delay, :logger, :limit_ram, :tmp_directory
 
     def initialize
       @storage_mode   = :local
@@ -17,6 +17,7 @@ module DataDrain
       @batch_size     = 5000
       @throttle_delay = 0.5
       @limit_ram      = nil # eg 2GB
+      @tmp_directory  = nil # eg /tmp/duckdb_work
       @logger         = Logger.new($stdout)
     end
 

data/lib/data_drain/engine.rb

@@ -81,6 +81,8 @@ module DataDrain
     def setup_duckdb
       @duckdb.query("INSTALL postgres; LOAD postgres;")
       @duckdb.query("SET max_memory='#{@config.limit_ram}';") if @config.limit_ram.present?
+      @duckdb.query("SET temp_directory='#{@config.tmp_directory}'") if @config.tmp_directory.present?
+      @duckdb.query("ATTACH '#{@config.duckdb_connection_string}' AS pg_source (TYPE POSTGRES, READ_ONLY)")
 
       # 💡 Magia del Adapter: Él sabe si cargar httpfs y setear credenciales o no hacer nada
       @adapter.setup_duckdb(@duckdb)
@@ -91,7 +93,7 @@ module DataDrain
     def get_postgres_count
       pg_sql = "SELECT COUNT(*) AS row_count FROM public.#{@table_name} WHERE #{base_where_sql}"
       pg_sql = pg_sql.gsub("'", "''")
-      query = "SELECT row_count FROM postgres_query('#{@config.duckdb_connection_string}', '#{pg_sql}')"
+      query = "SELECT row_count FROM postgres_query('pg_source', '#{pg_sql}')"
       @duckdb.query(query).first.first
     end
 
@@ -109,7 +111,7 @@ module DataDrain
       query = <<~SQL
         COPY (
           SELECT #{@select_sql}
-          FROM postgres_query('#{@config.duckdb_connection_string}', '#{pg_sql}')
+          FROM postgres_query('pg_source', '#{pg_sql}')
         ) TO '#{dest_path}'
         (
           FORMAT PARQUET,

data/lib/data_drain/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module DataDrain
-  VERSION = "0.1.4"
+  VERSION = "0.1.6"
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: data_drain
 version: !ruby/object:Gem::Version
-  version: 0.1.4
+  version: 0.1.6
 platform: ruby
 authors:
 - Gabriel