hivemeta 0.1.1 → 0.1.9

data/CHANGELOG

@@ -1,3 +1,11 @@
+* 2011-07-06 - fsf
+- new: added introductory jruby/jdbc support
+- change: HiveMeta::Connection.initialize takes 4 optional arguments now
+          and builds the dbi/jdbc string itself
+
+* 2011-05-30 - fsf
+- clean: switched dbi connection loop to an exception retry block
+
 * 2011-05-29 - fsf
 - new: added retry logic to the db connection to account for mysql servers
        with a relatively low max_connections setting

data/README

@@ -39,6 +39,10 @@ Installation
 
 gem install hivemeta
 
+or (for jruby)
+
+jgem install hivemeta
+
 ---
 
 API Usage
@@ -81,8 +85,7 @@ db_pass    = 'hivepasshere'
 db_host    = 'localhost'
 db_name    = 'hivemeta'
 
-dbi_string = "DBI:Mysql:#{db_name}:#{db_host}"
-h = HiveMeta::Connection.new(dbi_string, db_user, db_pass)
+h = HiveMeta::Connection.new(db_name, db_host, db_user, db_pass)
 
 establishing a connection (environment variables):
 

data/examples/hivemeta_query.rb

@@ -72,8 +72,7 @@ opts.each do |opt, arg|
       exit
   end end
 
-dbi_string = "DBI:Mysql:#{db_name}:#{db_host}"
-h = HiveMeta::Connection.new(dbi_string, db_user, db_pass)
+h = HiveMeta::Connection.new(db_name, db_host, db_user, db_pass)
 
 tables = []
 max_col_width = 8

data/examples/hivemeta_testrec.rb

@@ -44,8 +44,7 @@ opts.each do |opt, arg|
       exit
   end end
 
-dbi_string = "DBI:Mysql:#{db_name}:#{db_host}"
-h = HiveMeta::Connection.new(dbi_string, db_user, db_pass)
+h = HiveMeta::Connection.new(db_name, db_host, db_user, db_pass)
 
 # test table has the following schema
 # i   col_name

data/examples/sample-mapper.rb

@@ -10,8 +10,7 @@ db_pass    = 'hivepasshere'
 db_host    = 'localhost'
 db_name    = 'hivemeta'
 
-dbi_string = "DBI:Mysql:#{db_name}:#{db_host}"
-h = HiveMeta::Connection.new(dbi_string, db_user, db_pass)
+h = HiveMeta::Connection.new(db_name, db_host, db_user, db_pass)
 
 inv_table = h.table 'sample_inventory'
 

data/lib/hivemeta/connection.rb

@@ -1,48 +1,58 @@
-require 'dbi'
 require 'hivemeta/table'
 require 'hivemeta/record'
 
+if RUBY_PLATFORM == 'java'
+  require 'java'
+else
+  require 'dbi'
+  # fix for broken row dup in 1.9
+  # http://goo.gl/fx6kW
+  module DBI
+    class Row
+      if RUBY_VERSION =~ /^1\.9/
+        def dup
+          row = super
+          row.instance_variable_set :@arr, @arr.dup
+          row
+        end
+      end
+    end
+  end
+end
+
 module HiveMeta
 
   class Connection
-    def initialize(dbi_string = nil, db_user = nil, db_pass = nil)
-      db_name = ENV['hivemeta_db_name']
-      db_host = ENV['hivemeta_db_host']
-      dbi_string ||= "DBI:Mysql:#{db_name}:#{db_host}"
-      db_user    ||= ENV['hivemeta_db_user']
-      db_pass    ||= ENV['hivemeta_db_pass']
-
-      @dbi_string = dbi_string
-      @db_user    = db_user
-      @db_pass    = db_pass
+    def initialize(db_name = nil, db_host = nil, db_user = nil, db_pass = nil)
+      @db_name = db_name || ENV['hivemeta_db_name']
+      @db_host = db_host || ENV['hivemeta_db_host']
+      @db_user = db_user || ENV['hivemeta_db_user']
+      @db_pass = db_pass || ENV['hivemeta_db_pass']
     end
 
-    def query(sql, *args)
+    def query_dbi(sql, *args)
       dbh = results = nil
+      dbi_string = "DBI:Mysql:#{@db_name}:#{@db_host}"
 
       # make a few attempts in the event that mysql has not been
       # configured with enough connections to handle many mappers
-      max_attempts = 3
-      1.upto(max_attempts) do |attempt|
-        begin
-          dbh = DBI.connect(@dbi_string, @db_user, @db_pass)
-          break
-        rescue DBI::DatabaseError => e
-          if attempt < max_attempts
-            attempt += 1
-            s = rand + 0.50
-            STDERR.puts "retrying hivemeta connection after %f seconds" % s
-            sleep s
-          else
-            STDERR.puts "cannot connect to metastore %s:\n  error (%s) %s" %
-              [@dbi_string, e.err, e.errstr]
-            raise
-          end
+      attempts, max_attempts = 0, 3
+      begin
+        dbh = DBI.connect(dbi_string, @db_user, @db_pass)
+      rescue DBI::DatabaseError => e
+        attempts += 1
+        if attempts < max_attempts
+          s = rand + 0.50
+          STDERR.puts "retrying hivemeta connection after %f seconds..." % s
+          sleep s
+          retry
+        else
+          warn "cannot connect to metastore on %s:\n  error %s\n  %s" %
+            [@db_host, e.err, e.errstr]
+          raise
         end
       end
 
-#puts "sql: #{sql}"
-#puts "args: #{args}"
       sth = dbh.prepare(sql)
       sth.execute(*args)
       if block_given?
@@ -58,6 +68,63 @@ module HiveMeta
       results # returns nil if a block is given
     end
 
+    def table_info_jdbc result
+      meta = result.meta_data
+      cols = meta.column_count
+      colnames = []
+      cols.times do |i|
+        colnames[i] = meta.column_name i+1
+      end
+      [cols, colnames]
+    end
+
+    def query_jdbc(sql, *args)
+      results = []
+      db_url = "jdbc:mysql://#{@db_host}/#{@db_name}"
+
+      # make a few attempts in the event that mysql has not been
+      # configured with enough connections to handle many mappers
+      attempts, max_attempts = 0, 3
+      begin
+        c = java.sql.DriverManager.get_connection(db_url, @db_user, @db_pass)
+      rescue => e
+        attempts += 1
+        if attempts < max_attempts
+          s = rand + 0.50
+          STDERR.puts "retrying hivemeta connection after %f seconds..." % s
+          sleep s
+          retry
+        else
+          warn "cannot connect to metastore on %s:\n  error %s" %
+            [@db_host, e]
+          raise
+        end
+      end
+
+      stmt = c.create_statement
+
+      args.each do |arg|
+        # poor man's prepare
+        sql = sql.sub /\?/, "'#{arg}'"
+        res = stmt.execute_query sql
+
+        cols,names = table_info_jdbc res
+
+        while res.next do 
+          row = []
+          1.upto(cols) do |i|
+            row << res.get_string(i)
+          end
+          yield row if block_given?
+          results << row
+        end
+      end
+
+      c.close
+
+      results
+    end
+
     def tables(opts = {})
       args = nil
       if opts[:filter_path]
@@ -76,13 +143,8 @@ module HiveMeta
       results = query sql, *args
       table_names = results.map {|result| result[0]}
       
-#puts "TABLE_NAMES:"
-#p table_names
-
       tables = []
       table_names.each do |name|
-#puts "NAME: "
-#p name
         table = Table.new(name)
 
         sql = "select c.INTEGER_IDX, c.column_name, c.COMMENT,
@@ -91,8 +153,6 @@ module HiveMeta
           where t.SD_ID = c.SD_ID and t.SD_ID = s.SD_ID
           and t.TBL_NAME = ?"
         query sql, name do |rec|
-#puts "REC:"
-#p rec
           col_idx  = rec[0].to_i
           col_name = rec[1]
           col_cmt  = rec[2]
@@ -113,8 +173,6 @@ module HiveMeta
         if results and results[0] and results[0][0]
           table.delimiter = results[0][0]
         end
-#puts "#{name}: found delim '#{table.delimiter}'" if results[0]
-#puts "#{name}: no delim" if not results[0]
 
         tables << table
       end
@@ -125,20 +183,13 @@ module HiveMeta
       t = tables(:filter_name => name) # appeasing the old skool 1.8 users
       t[0] # if it comes back with multiple tables, return the first
     end
-  end
 
-end
-
-# fix for broken row dup in 1.9
-# http://rubyforge.org/tracker/index.php?func=detail&aid=28624&group_id=234&atid=967
-module DBI
-  class Row
-    if RUBY_VERSION =~ /^1\.9/
-      def dup
-        row = super
-        row.instance_variable_set :@arr, @arr.dup
-        row
-      end
+    if RUBY_PLATFORM == 'java'
+      alias :query :query_jdbc
+      Java::com.mysql.jdbc.Driver
+    else
+      alias :query :query_dbi
     end
   end
+
 end

metadata

@@ -5,8 +5,8 @@ version: !ruby/object:Gem::Version
   segments: 
   - 0
   - 1
-  - 1
-  version: 0.1.1
+  - 9
+  version: 0.1.9
 platform: ruby
 authors: 
 - Frank Fejes
@@ -14,7 +14,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2011-05-29 00:00:00 -05:00
+date: 2011-07-06 00:00:00 -05:00
 default_executable: 
 dependencies: []