mobilize-hive 1.2 → 1.3

data/README.md

@@ -142,6 +142,17 @@ Start
   * cluster and user are optional for all of the below.
     * cluster defaults to the first cluster listed;
     * user is treated the same way as in [mobilize-ssh][mobilize-ssh].
+  * params are also optional for all of the below. They replace HQL in sources.
+    * params are passed as a YML or JSON, as in:
+      * `hive.run source:<source_path>, params:{'date': '2013-03-01', 'unit': 'widgets'}`
+        * this example replaces all the keys, preceded by '@' in all source hqls with the value.
+          * The preceding '@' is used to keep from replacing instances
+            of "date" and "unit" in the HQL; you should have `@date` and `@unit` in your actual HQL 
+            if you'd like to replace those tokens.
+    * in addition, the following params are substituted automatically:
+      * `$utc_date` - replaced with YYYY-MM-DD date, UTC
+      * `$utc_time` - replaced with HH:MM time, UTC
+      * any occurrence of these values in HQL will be replaced at runtime.
   * hive.run `hql:<hql> || source:<gsheet_path>, user:<user>, cluster:<cluster>`, which executes the
       script in the hql or source sheet and returns any output specified at the
       end. If the cmd or last query in source is a select statement, column headers will be

data/lib/mobilize-hive/handlers/hive.rb

@@ -1,58 +1,9 @@
 module Mobilize
   module Hive
-    def Hive.config
-      Base.config('hive')
-    end
-
-    def Hive.exec_path(cluster)
-      Hive.clusters[cluster]['exec_path']
-    end
-
-    def Hive.output_db(cluster)
-      Hive.clusters[cluster]['output_db']
-    end
-
-    def Hive.output_db_user(cluster)
-      output_db_node = Hadoop.gateway_node(cluster)
-      output_db_user = Ssh.host(output_db_node)['user']
-      output_db_user
-    end
-
-    def Hive.clusters
-      Hive.config['clusters']
-    end
-
-    def Hive.slot_ids(cluster)
-      (1..Hive.clusters[cluster]['max_slots']).to_a.map{|s| "#{cluster}_#{s.to_s}"}
-    end
-
-    def Hive.slot_worker_by_cluster_and_path(cluster,path)
-      working_slots = Mobilize::Resque.jobs.map{|j| begin j['args'][1]['hive_slot'];rescue;nil;end}.compact.uniq
-      Hive.slot_ids(cluster).each do |slot_id|
-        unless working_slots.include?(slot_id)
-          Mobilize::Resque.set_worker_args_by_path(path,{'hive_slot'=>slot_id})
-          return slot_id
-        end
-      end
-      #return false if none are available
-      return false
-    end
-
-    def Hive.unslot_worker_by_path(path)
-      begin
-        Mobilize::Resque.set_worker_args_by_path(path,{'hive_slot'=>nil})
-        return true
-      rescue
-        return false
-      end
-    end
-
-    def Hive.databases(cluster,user_name)
-      Hive.run(cluster,"show databases",user_name)['stdout'].split("\n")
-    end
-
+    #adds convenience methods
+    require "#{File.dirname(__FILE__)}/../helpers/hive_helper"
     # converts a source path or target path to a dst in the context of handler and stage
-    def Hive.path_to_dst(path,stage_path)
+    def Hive.path_to_dst(path,stage_path,gdrive_slot)
       has_handler = true if path.index("://")
       s = Stage.where(:path=>stage_path).first
       params = s.params
@@ -78,7 +29,7 @@ module Mobilize
         return Dataset.find_or_create_by_url(hive_url)
       end
       #otherwise, use hdfs convention
-      return Ssh.path_to_dst(path,stage_path)
+      return Ssh.path_to_dst(path,stage_path,gdrive_slot)
     end
 
     def Hive.url_by_path(path,user_name,is_target=false)
@@ -108,7 +59,7 @@ module Mobilize
     def Hive.table_stats(cluster,db,table,user_name)
       describe_sql = "use #{db};describe extended #{table};"
       describe_response = Hive.run(cluster, describe_sql,user_name)
-      return describe_response if describe_response['stdout'].length==0
+      return nil if describe_response['stdout'].length==0
       describe_output = describe_response['stdout']
       describe_output.split("location:").last.split(",").first
       #get location, fields, partitions
@@ -142,20 +93,43 @@ module Mobilize
     end
 
     #run a generic hive command, with the option of passing a file hash to be locally available
-    def Hive.run(cluster,hql,user_name,file_hash=nil)
+    def Hive.run(cluster,hql,user_name,params=nil,file_hash=nil)
       # no TempStatsStore
       hql = "set hive.stats.autogather=false;#{hql}"
       filename = hql.to_md5
       file_hash||= {}
       file_hash[filename] = hql
+      #add in default params
+      params ||= {}
+      params = params.merge(Hive.default_params)
+      #replace any params in the file_hash and command
+      params.each do |k,v|
+        file_hash.each do |name,data|
+          if k.starts_with?("$")
+            data.gsub!(k,v)
+          else
+            data.gsub!("@#{k}",v)
+          end
+        end
+      end
       #silent mode so we don't have logs in stderr; clip output
       #at hadoop read limit
       command = "#{Hive.exec_path(cluster)} -S -f #{filename} | head -c #{Hadoop.read_limit}"
       gateway_node = Hadoop.gateway_node(cluster)
-      Ssh.run(gateway_node,command,user_name,file_hash)
+      response = Ssh.run(gateway_node,command,user_name,file_hash)
+      #override exit code 0 when stdout is blank and
+      #stderror contains FAILED or KILLED
+      if response['stdout'].to_s.length == 0 and
+        response['stderr'].to_s.ie{|se| se.index("FAILED") or se.index("KILLED")}
+        response['exit_code'] = 500
+      end
+      return response
     end
 
     def Hive.run_by_stage_path(stage_path)
+      gdrive_slot = Gdrive.slot_worker_by_path(stage_path)
+      #return blank response if there are no slots available
+      return nil unless gdrive_slot
       s = Stage.where(:path=>stage_path).first
       params = s.params
       cluster = params['cluster'] || Hive.clusters.keys.first
@@ -174,13 +148,16 @@ module Mobilize
       if params['hql']
         hql = params['hql']
       else
-        source = s.sources.first
-        hql = source.read(user_name)
+        source = s.sources(gdrive_slot).first
+        hql = source.read(user_name,gdrive_slot)
       end
 
+      Gdrive.unslot_worker_by_path(stage_path)
+
       #check for select at end
       hql_array = hql.split(";").map{|hc| hc.strip}.reject{|hc| hc.length==0}
-      if hql_array.last.downcase.starts_with?("select")
+      last_statement = hql_array.last.downcase.split("\n").reject{|l| l.starts_with?("-- ")}.first
+      if last_statement.to_s.starts_with?("select")
         #nil if no prior commands
         prior_hql = hql_array[0..-2].join(";") if hql_array.length > 1
         select_hql = hql_array.last
@@ -188,10 +165,10 @@ module Mobilize
                             "drop table if exists #{output_path}",
                             "create table #{output_path} as #{select_hql};"].join(";")
         full_hql = [prior_hql, output_table_hql].compact.join(";")
-        result = Hive.run(cluster,full_hql, user_name)
+        result = Hive.run(cluster,full_hql, user_name,params['params'])
         Dataset.find_or_create_by_url(out_url)
       else
-        result = Hive.run(cluster, hql, user_name)
+        result = Hive.run(cluster, hql, user_name,params['params'])
         Dataset.find_or_create_by_url(out_url)
         Dataset.write_by_url(out_url,result['stdout'],user_name) if result['stdout'].to_s.length>0
       end
@@ -224,7 +201,7 @@ module Mobilize
       file_name = schema_path.split("/").last
       out_url = "gridfs://#{schema_path}/#{file_name}"
       Dataset.write_by_url(out_url,out_tsv,user_name)
-      schema_tsv = Dataset.find_by_url(out_url).read(user_name)
+      schema_tsv = Dataset.find_by_url(out_url).read(user_name,gdrive_slot)
       schema_hash = {}
       schema_tsv.tsv_to_hash_array.each do |ha|
         schema_hash[ha['name']] = ha['datatype']
@@ -232,24 +209,10 @@ module Mobilize
       schema_hash
     end
 
-    def Hive.path_params(cluster, path, user_name)
-      db, table, partitions = path.gsub(".","/").split("/").ie{|sp| [sp.first, sp.second, sp[2..-1]]}
-      #get existing table stats if any
-      curr_stats = begin
-                     Hive.table_stats(cluster, db, table, user_name)
-                   rescue
-                     nil
-                   end
-      {"db"=>db,
-       "table"=>table,
-       "partitions"=>partitions,
-       "curr_stats"=>curr_stats}
-    end
-
-    def Hive.hql_to_table(cluster, db, table, part_array, source_hql, user_name, job_name, drop=false, schema_hash=nil)
+    def Hive.hql_to_table(cluster, db, table, part_array, source_hql, user_name, job_name, drop=false, schema_hash=nil, params=nil)
       table_path = [db,table].join(".")
-      target_params = Hive.path_params(cluster, table_path, user_name)
-      table_stats = target_params['curr_stats']
+      table_stats = Hive.table_stats(cluster, db, table, user_name)
+      url = "hive://" + [cluster,db,table,part_array.compact.join("/")].join("/")
 
       source_hql_array = source_hql.split(";")
       last_select_i = source_hql_array.rindex{|hql| hql.downcase.strip.starts_with?("select")}
@@ -265,11 +228,10 @@ module Mobilize
       temp_set_hql = "set mapred.job.name=#{job_name} (temp table);"
       temp_drop_hql = "drop table if exists #{temp_table_path};"
       temp_create_hql = "#{temp_set_hql}#{prior_hql}#{temp_drop_hql}create table #{temp_table_path} as #{last_select_hql}"
-      Hive.run(cluster,temp_create_hql,user_name)
+      response = Hive.run(cluster,temp_create_hql,user_name,params)
+      raise response['stderr'] if response['stderr'].to_s.ie{|s| s.index("FAILED") or s.index("KILLED")}
 
-      source_params = Hive.path_params(cluster, temp_table_path, user_name)
-      source_table_path = ['db','table'].map{|k| source_params[k]}.join(".")
-      source_table_stats = source_params['curr_stats']
+      source_table_stats = Hive.table_stats(cluster,temp_db,temp_table_name,user_name)
       source_fields = source_table_stats['field_defs']
 
       if part_array.length == 0 and
@@ -297,7 +259,7 @@ module Mobilize
 
         target_create_hql = "create table if not exists #{table_path} #{field_def_stmt};"
 
-        target_insert_hql = "insert overwrite table #{table_path} select #{target_field_stmt} from #{source_table_path};"
+        target_insert_hql = "insert overwrite table #{table_path} select #{target_field_stmt} from #{temp_table_path};"
 
         target_full_hql = [target_name_hql,
                            target_drop_hql,
@@ -305,10 +267,12 @@ module Mobilize
                            target_insert_hql,
                            temp_drop_hql].join
 
-        Hive.run(cluster, target_full_hql, user_name)
+        response = Hive.run(cluster, target_full_hql, user_name, params)
+
+        raise response['stderr'] if response['stderr'].to_s.ie{|s| s.index("FAILED") or s.index("KILLED")}
 
       elsif part_array.length > 0 and
-        table_stats.ie{|tts| tts.nil? || drop || tts['partitions'].to_a.map{|p| p['name']} == part_array}
+        table_stats.ie{|tts| tts.nil? || drop || tts['partitions'].to_a.map{|p| p['name']}.sort == part_array.sort}
         #partitions and no target table or same partitions in both target table and user params
 
         target_headers = source_fields.map{|f| f['name']}.reject{|h| part_array.include?(h)}
@@ -352,10 +316,20 @@ module Mobilize
 
         else
           #get all the permutations of possible partititons
-          part_perm_hql = "set hive.cli.print.header=true;select distinct #{target_part_stmt} from #{source_table_path};"
-          part_perm_tsv = Hive.run(cluster, part_perm_hql, user_name)['stdout']
+          part_set_hql = "set hive.cli.print.header=true;set mapred.job.name=#{job_name} (permutations);"
+          part_select_hql = "select distinct #{target_part_stmt} from #{temp_table_path};"
+          part_perm_hql = part_set_hql + part_select_hql
+          response = Hive.run(cluster, part_perm_hql, user_name, params)
+          raise response['stderr'] if response['stderr'].to_s.ie{|s| s.index("FAILED") or s.index("KILLED")}
+          part_perm_tsv = response['stdout']
           #having gotten the permutations, ensure they are dropped
           part_hash_array = part_perm_tsv.tsv_to_hash_array
+          #make sure there is data
+          if part_hash_array.first.nil? or part_hash_array.first.values.include?(nil)
+            #blank result set, return url
+            return url
+          end
+
           part_drop_hql = part_hash_array.map do |h|
             part_drop_stmt = h.map do |name,value|
                                part_defs[name[1..-2]]=="string" ? "#{name}='#{value}'" : "#{name}=#{value}"
@@ -367,16 +341,16 @@ module Mobilize
 
         target_insert_hql = "insert overwrite table #{table_path} " +
                             "partition (#{target_part_stmt}) " +
-                            "select #{target_field_stmt},#{target_part_stmt} from #{source_table_path};"
+                            "select #{target_field_stmt},#{target_part_stmt} from #{temp_table_path};"
 
         target_full_hql = [target_set_hql, target_create_hql, target_insert_hql, temp_drop_hql].join
 
-        Hive.run(cluster, target_full_hql, user_name)
+        response = Hive.run(cluster, target_full_hql, user_name, params)
+        raise response['stderr'] if response['stderr'].to_s.ie{|s| s.index("FAILED") or s.index("KILLED")}
       else
         error_msg = "Incompatible partition specs"
         raise error_msg
       end
-      url = "hive://" + [cluster,db,table,part_array.compact.join("/")].join("/")
       return url
     end
 
@@ -384,14 +358,21 @@ module Mobilize
     #Accepts options to drop existing target if any
     #also schema with column datatype overrides
     def Hive.tsv_to_table(cluster, db, table, part_array, source_tsv, user_name, drop=false, schema_hash=nil)
+      return nil if source_tsv.strip.length==0
+      if source_tsv.index("\r\n")
+        source_tsv = source_tsv.gsub("\r\n","\n")
+      elsif source_tsv.index("\r")
+        source_tsv = source_tsv.gsub("\r","\n")
+      end
       source_headers = source_tsv.tsv_header_array
 
       table_path = [db,table].join(".")
-      target_params = Hive.path_params(cluster, table_path, user_name)
-      table_stats = target_params['curr_stats']
+      table_stats = Hive.table_stats(cluster, db, table, user_name)
 
       schema_hash ||= {}
 
+      url = "hive://" + [cluster,db,table,part_array.compact.join("/")].join("/")
+
       if part_array.length == 0 and
         table_stats.ie{|tts| tts.nil? || drop || tts['partitions'].nil?}
         #no partitions in either user params or the target table
@@ -418,10 +399,11 @@ module Mobilize
 
         target_full_hql = [target_drop_hql,target_create_hql,target_insert_hql].join(";")
 
-        Hive.run(cluster, target_full_hql, user_name, file_hash)
+        response = Hive.run(cluster, target_full_hql, user_name, nil, file_hash)
+        raise response['stderr'] if response['stderr'].to_s.ie{|s| s.index("FAILED") or s.index("KILLED")}
 
       elsif part_array.length > 0 and
-        table_stats.ie{|tts| tts.nil? || drop || tts['partitions'].to_a.map{|p| p['name']} == part_array}
+        table_stats.ie{|tts| tts.nil? || drop || tts['partitions'].to_a.map{|p| p['name']}.sort == part_array.sort}
         #partitions and no target table
         #or same partitions in both target table and user params
         #or drop and start fresh
@@ -445,13 +427,17 @@ module Mobilize
                             "partitioned by #{partition_defs}"
 
         #create target table early if not here
-        Hive.run(cluster, target_create_hql, user_name)
+        response = Hive.run(cluster, target_create_hql, user_name)
+        raise response['stderr'] if response['stderr'].to_s.ie{|s| s.index("FAILED") or s.index("KILLED")}
+
+        #return url (operation complete) if there's no data
+        source_hash_array = source_tsv.tsv_to_hash_array
+        return url if source_hash_array.length==1 and source_hash_array.first.values.compact.length==0
 
         table_stats = Hive.table_stats(cluster, db, table, user_name)
 
         #create data hash from source hash array
         data_hash = {}
-        source_hash_array = source_tsv.tsv_to_hash_array
         source_hash_array.each do |ha|
           tpmk = part_array.map{|pn| "#{pn}=#{ha[pn]}"}.join("/")
           tpmv = ha.reject{|k,v| part_array.include?(k)}.values.join("\001")
@@ -484,7 +470,8 @@ module Mobilize
         #run actual partition adds all at once
         if target_part_hql.length>0
           puts "Adding partitions to #{cluster}/#{db}/#{table} for #{user_name} at #{Time.now.utc}"
-          Hive.run(cluster, target_part_hql, user_name)
+          response = Hive.run(cluster, target_part_hql, user_name)
+          raise response['stderr'] if response['stderr'].to_s.ie{|s| s.index("FAILED") or s.index("KILLED")}
         end
       else
         error_msg = "Incompatible partition specs: " +
@@ -492,33 +479,31 @@ module Mobilize
                     "user_params:#{part_array.to_s}"
         raise error_msg
       end
-      url = "hive://" + [cluster,db,table,part_array.compact.join("/")].join("/")
+
       return url
     end
 
     def Hive.write_by_stage_path(stage_path)
+      gdrive_slot = Gdrive.slot_worker_by_path(stage_path)
+      #return blank response if there are no slots available
+      return nil unless gdrive_slot
       s = Stage.where(:path=>stage_path).first
       params = s.params
-      source = s.sources.first
+      source = s.sources(gdrive_slot).first
       target = s.target
       cluster, db, table = target.url.split("://").last.split("/")
-      #update stage with the node so we can use it
-      user_name = Hdfs.user_name_by_stage_path(stage_path,cluster)
-      job_name = s.path.sub("Runner_","")
-
       #slot Hive worker if available
       slot_id = Hive.slot_worker_by_cluster_and_path(cluster,stage_path)
       return false unless slot_id
+      #update stage with the node so we can use it
+      user_name = Hdfs.user_name_by_stage_path(stage_path,cluster)
+      job_name = s.path.sub("Runner_","")
 
       schema_hash = if params['schema']
-                      gdrive_slot = Gdrive.slot_worker_by_path(stage_path)
-                      #return blank response if there are no slots available
-                      return nil unless gdrive_slot
                       Hive.schema_hash(params['schema'],user_name,gdrive_slot)
                     else
                       {}
                     end
-      Gdrive.unslot_worker_by_path(stage_path)
       #drop target before create/insert?
       drop = params['drop']
 
@@ -531,16 +516,17 @@ module Mobilize
           #source table
           cluster,source_path = source.path.split("/").ie{|sp| [sp.first, sp[1..-1].join(".")]}
           source_hql = "select * from #{source_path};"
-        elsif ['gsheet','gridfs','hdfs'].include?(source.handler)
+        elsif ['gsheet','gfile','gridfs','hdfs'].include?(source.handler)
           if source.path.ie{|sdp| sdp.index(/\.[A-Za-z]ql$/) or sdp.ends_with?(".ql")}
-            source_hql = source.read(user_name)
+            source_hql = source.read(user_name,gdrive_slot)
           else
-            #tsv from sheet
-            source_tsv = source.read(user_name)
+            #tsv from sheet or file
+            source_tsv = source.read(user_name,gdrive_slot)
           end
         end
       end
 
+      Gdrive.unslot_worker_by_path(stage_path)
       part_array = if params['partitions']
                     params['partitions'].to_a.map{|p| p.gsub(".","/").split("/")}.flatten
                   elsif params['target']
@@ -559,12 +545,14 @@ module Mobilize
                          Hive.hql_to_table(cluster, db, table, part_array, source_hql, user_name, job_name, drop, schema_hash)
                        elsif source_tsv
                          Hive.tsv_to_table(cluster, db, table, part_array, source_tsv, user_name, drop, schema_hash)
+                       elsif source
+                         #null sheet
                        else
                          raise "Unable to determine source tsv or source hql"
                        end
                  {'stdout'=>url,'exit_code'=>0}
                rescue => exc
-                 {'stderr'=>exc.to_s, 'exit_code'=>500}
+                 {'stderr'=>"#{exc.to_s}\n#{exc.backtrace.join("\n")}", 'exit_code'=>500}
                end
 
       #unslot worker and write result
@@ -585,11 +573,8 @@ module Mobilize
       select_hql = "select * from #{source_path};"
       hql = [set_hql,select_hql].join
       response = Hive.run(cluster, hql,user_name)
-      if response['exit_code']==0
-        return response['stdout']
-      else
-        raise "Unable to read hive://#{dst_path} with error: #{response['stderr']}"
-      end
+      raise "Unable to read hive://#{dst_path} with error: #{response['stderr']}" if response['stderr'].to_s.ie{|s| s.index("FAILED") or s.index("KILLED")}
+      return response['stdout']
     end
 
     def Hive.write_by_dataset_path(dst_path,source_tsv,user_name,*args)

data/lib/mobilize-hive/helpers/hive_helper.rb

@@ -0,0 +1,63 @@
+module Mobilize
+  module Hive
+    def self.config
+      Base.config('hive')
+    end
+
+    def self.exec_path(cluster)
+      self.clusters[cluster]['exec_path']
+    end
+
+    def self.output_db(cluster)
+      self.clusters[cluster]['output_db']
+    end
+
+    def self.output_db_user(cluster)
+      output_db_node = Hadoop.gateway_node(cluster)
+      output_db_user = Ssh.host(output_db_node)['user']
+      output_db_user
+    end
+
+    def self.clusters
+      self.config['clusters']
+    end
+
+    def self.slot_ids(cluster)
+      (1..self.clusters[cluster]['max_slots']).to_a.map{|s| "#{cluster}_#{s.to_s}"}
+    end
+
+    def self.slot_worker_by_cluster_and_path(cluster,path)
+      working_slots = Mobilize::Resque.jobs.map{|j| begin j['args'][1]['hive_slot'];rescue;nil;end}.compact.uniq
+      self.slot_ids(cluster).each do |slot_id|
+        unless working_slots.include?(slot_id)
+          Mobilize::Resque.set_worker_args_by_path(path,{'hive_slot'=>slot_id})
+          return slot_id
+        end
+      end
+      #return false if none are available
+      return false
+    end
+
+    def self.unslot_worker_by_path(path)
+      begin
+        Mobilize::Resque.set_worker_args_by_path(path,{'hive_slot'=>nil})
+        return true
+      rescue
+        return false
+      end
+    end
+
+    def self.databases(cluster,user_name)
+      self.run(cluster,"show databases",user_name)['stdout'].split("\n")
+    end
+
+    def self.default_params
+      time = Time.now.utc
+      {
+       '$utc_date'=>time.strftime("%Y-%m-%d"),
+       '$utc_time'=>time.strftime("%H:%M"),
+      }
+    end
+  end
+end
+

data/lib/mobilize-hive/version.rb

@@ -1,5 +1,5 @@
 module Mobilize
   module Hive
-    VERSION = "1.2"
+    VERSION = "1.3"
   end
 end

data/mobilize-hive.gemspec

@@ -16,5 +16,5 @@ Gem::Specification.new do |gem|
   gem.executables   = gem.files.grep(%r{^bin/}).map{ |f| File.basename(f) }
   gem.test_files    = gem.files.grep(%r{^(test|spec|features)/})
   gem.require_paths = ["lib"]
-  gem.add_runtime_dependency "mobilize-hdfs","1.2"
+  gem.add_runtime_dependency "mobilize-hdfs","1.3"
 end

data/test/hive_job_rows.yml

@@ -20,7 +20,15 @@
   active: true
   trigger: after hive_test_2
   status: ""
-  stage1: hive.run hql:"select act_date as `date`,product,category,value from mobilize.hive_test_1;"
+  stage1: hive.run hql:"select '@date' as `date`,product,category,value from mobilize.hive_test_1;", params:{'date':'2013-01-01'}
   stage2: hive.write source:"stage1",target:"mobilize/hive_test_3", partitions:"date/product", drop:true
   stage3: hive.write hql:"select * from mobilize.hive_test_3;",target:"mobilize/hive_test_3", partitions:"date/product", drop:false
   stage4: gsheet.write source:"hive://mobilize/hive_test_3", target:"hive_test_3.out"
+- name: hive_test_4
+  active: true
+  trigger: after hive_test_3
+  status: ""
+  stage1: hive.write source:"hive_test_4_stage_1.in", target:"mobilize/hive_test_1", partitions:"act_date"
+  stage2: hive.write source:"hive_test_4_stage_2.in", target:"mobilize/hive_test_1", partitions:"act_date"
+  stage3: hive.run hql:"select '$utc_date $utc_time' as `date_time`,product,category,value from mobilize.hive_test_1;"
+  stage4: gsheet.write source:stage3, target:"hive_test_4.out"

data/test/mobilize-hive_test.rb

@@ -25,6 +25,18 @@ describe "Mobilize" do
     hive_1_in_tsv = YAML.load_file("#{Mobilize::Base.root}/test/hive_test_1_in.yml").hash_array_to_tsv
     hive_1_in_sheet.write(hive_1_in_tsv,Mobilize::Gdrive.owner_name)
 
+    #create blank sheet
+    hive_4_stage_1_in_sheet = Mobilize::Gsheet.find_by_path("#{r.path.split("/")[0..-2].join("/")}/hive_test_4_stage_1.in",gdrive_slot)
+    [hive_4_stage_1_in_sheet].each {|s| s.delete if s}
+    hive_4_stage_1_in_sheet = Mobilize::Gsheet.find_or_create_by_path("#{r.path.split("/")[0..-2].join("/")}/hive_test_4_stage_1.in",gdrive_slot)
+
+    #create sheet w just headers
+    hive_4_stage_2_in_sheet = Mobilize::Gsheet.find_by_path("#{r.path.split("/")[0..-2].join("/")}/hive_test_4_stage_2.in",gdrive_slot)
+    [hive_4_stage_2_in_sheet].each {|s| s.delete if s}
+    hive_4_stage_2_in_sheet = Mobilize::Gsheet.find_or_create_by_path("#{r.path.split("/")[0..-2].join("/")}/hive_test_4_stage_2.in",gdrive_slot)
+    hive_4_stage_2_in_sheet_header = hive_1_in_tsv.tsv_header_array.join("\t")
+    hive_4_stage_2_in_sheet.write(hive_4_stage_2_in_sheet_header,Mobilize::Gdrive.owner_name)
+
     hive_1_schema_sheet = Mobilize::Gsheet.find_by_path("#{r.path.split("/")[0..-2].join("/")}/hive_test_1.schema",gdrive_slot)
     [hive_1_schema_sheet].each {|s| s.delete if s}
     hive_1_schema_sheet = Mobilize::Gsheet.find_or_create_by_path("#{r.path.split("/")[0..-2].join("/")}/hive_test_1.schema",gdrive_slot)
@@ -51,21 +63,25 @@ describe "Mobilize" do
     [hive_2_target_sheet].each{|s| s.delete if s}
     hive_3_target_sheet = Mobilize::Gsheet.find_by_path("#{r.path.split("/")[0..-2].join("/")}/hive_test_3.out",gdrive_slot)
     [hive_3_target_sheet].each{|s| s.delete if s}
+    hive_4_target_sheet = Mobilize::Gsheet.find_by_path("#{r.path.split("/")[0..-2].join("/")}/hive_test_4.out",gdrive_slot)
+    [hive_4_target_sheet].each{|s| s.delete if s}
 
     puts "job row added, force enqueued requestor, wait for stages"
     r.enqueue!
-    wait_for_stages(1200)
+    wait_for_stages(2100)
 
     puts "jobtracker posted data to test sheet"
     hive_1_stage_2_target_sheet = Mobilize::Gsheet.find_by_path("#{r.path.split("/")[0..-2].join("/")}/hive_test_1_stage_2.out",gdrive_slot)
     hive_1_stage_3_target_sheet = Mobilize::Gsheet.find_by_path("#{r.path.split("/")[0..-2].join("/")}/hive_test_1_stage_3.out",gdrive_slot)
     hive_2_target_sheet = Mobilize::Gsheet.find_by_path("#{r.path.split("/")[0..-2].join("/")}/hive_test_2.out",gdrive_slot)
     hive_3_target_sheet = Mobilize::Gsheet.find_by_path("#{r.path.split("/")[0..-2].join("/")}/hive_test_3.out",gdrive_slot)
+    hive_4_target_sheet = Mobilize::Gsheet.find_by_path("#{r.path.split("/")[0..-2].join("/")}/hive_test_4.out",gdrive_slot)
 
     assert hive_1_stage_2_target_sheet.read(u.name).length == 219
     assert hive_1_stage_3_target_sheet.read(u.name).length > 3
     assert hive_2_target_sheet.read(u.name).length == 599
     assert hive_3_target_sheet.read(u.name).length == 347
+    assert hive_4_target_sheet.read(u.name).length == 432
   end
 
   def wait_for_stages(time_limit=600,stage_limit=120,wait_length=10)

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: mobilize-hive
 version: !ruby/object:Gem::Version
-  version: '1.2'
+  version: '1.3'
   prerelease: 
 platform: ruby
 authors:
@@ -9,7 +9,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2013-03-21 00:00:00.000000000 Z
+date: 2013-04-18 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: mobilize-hdfs
@@ -18,7 +18,7 @@ dependencies:
     requirements:
     - - '='
       - !ruby/object:Gem::Version
-        version: '1.2'
+        version: '1.3'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
@@ -26,7 +26,7 @@ dependencies:
     requirements:
     - - '='
       - !ruby/object:Gem::Version
-        version: '1.2'
+        version: '1.3'
 description: Adds hive read, write, and run support to mobilize-hdfs
 email:
 - cpaesleme@dena.com
@@ -41,6 +41,7 @@ files:
 - Rakefile
 - lib/mobilize-hive.rb
 - lib/mobilize-hive/handlers/hive.rb
+- lib/mobilize-hive/helpers/hive_helper.rb
 - lib/mobilize-hive/tasks.rb
 - lib/mobilize-hive/version.rb
 - lib/samples/hive.yml
@@ -72,7 +73,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
       version: '0'
 requirements: []
 rubyforge_project: 
-rubygems_version: 1.8.24
+rubygems_version: 1.8.25
 signing_key: 
 specification_version: 3
 summary: Adds hive read, write, and run support to mobilize-hdfs
@@ -84,4 +85,3 @@ test_files:
 - test/mobilize-hive_test.rb
 - test/redis-test.conf
 - test/test_helper.rb
-has_rdoc: