hivemeta 0.0.2 → 0.0.3

data/README

@@ -1,15 +1,47 @@
 hivemeta
 
-A ruby API for access to the Hive metastore.  Useful for querying columns
-in Hadoop map/reduce applications.  Includes a demo application to spit out
-table information from the command-line via table name search or
-by the table's location in HDFS.
+A ruby API for access to a Hive metastore running under MySQL.
+
+Useful for querying columns in Hadoop map/reduce applications.  Normally,
+a developer needs to handle both the splitting of incoming data and the
+assignment of numerically indexed fields to friendly variables like so:
+
+fields = line.chomp.split /\t/
+item_id = fields[0]
+inv_cnt = fields[7].to_i
+puts "#{item_id}\t#{inv_cnt}"
+
+This is not overly traumatic, however it's susceptible to errors creeping
+in from file format changes.  Ongoing maintenance can easily become a burden
+if there are many map/reduce programs reading the same changed data files.
+Code size increases as the column count increases.
+
+With hivemeta, the process is streamlined.  That same task is now:
+
+row = inv_table.process_row line
+puts "#{row.item_id}\t#{row.inv_cnt.to_i}"
+
+The row object automagically knows its column names and they can be
+referenced in one of the following ways (in order of best to worst
+performance and coolness):
+
+row.col_name
+row[:col_name]
+row['col_name']
+
+Also included is a demo application, hivemeta_query.rb, to spit out table
+information from the command-line via table name search or by the table's
+location in HDFS.
+
+---
 
 Installation
 
 gem install hivemeta
 
-Usage
+---
+
+API Usage
 
 streaming map/reduce code snippet:
 
@@ -30,12 +62,15 @@ STDIN.each_line do |line|
   puts "#{item_id}\t#{count}" if count >= 1000
 end
 
-sample usage for the demo app:
+---
+
+hivemeta_query.rb Usage
 
 # query by table names
 $ hivemeta_query.rb join_test_name
 join_test_name
 hdfs://namenode/tmp/join_test_name
+delimiter: "\t" (ASCII 9)
 0   userid             # userid
 1   name               # username
 
@@ -43,6 +78,7 @@ hdfs://namenode/tmp/join_test_name
 $ hivemeta_query.rb join_test%
 join_test_address
 hdfs://namenode/tmp/join_test_address
+delimiter: "," (ASCII 44)
 0   userid             # uid
 1   address
 2   city
@@ -50,6 +86,7 @@ hdfs://namenode/tmp/join_test_address
 
 join_test_name
 hdfs://namenode/tmp/join_test_name
+delimiter: "\t" (ASCII 9)
 0   userid             # userid
 1   name               # username
 

data/lib/hivemeta/connection.rb

@@ -80,7 +80,6 @@ module HiveMeta
           table.columns[col_idx]  = col_name
           table.comments[col_idx] = col_cmt
           table.path      = tbl_loc
-
         end
 
         sql = "select sp.PARAM_VALUE

data/lib/hivemeta/record.rb

@@ -4,22 +4,30 @@ module HiveMeta
 
   class Record
     def initialize(line, table)
-      fields = line.chomp.split(table.delimiter, -1)
-      if fields.size != table.columns.size
+      @fields = line.chomp.split(table.delimiter, -1)
+      if @fields.size != table.columns.size
         raise FieldCountError
       end
 
       @columns = {}
       table.each_col_with_index do |col_name, i|
-        @columns[col_name] = fields[i]
-        @columns[col_name.to_sym] = fields[i]
+        @columns[col_name] = @fields[i]
+        @columns[col_name.to_sym] = @fields[i]
       end
     end
 
+    # allow for column access via column name as an index
+    # example: rec[:col_name]
+    #      or: rec['col_name']
+    # can also use the numeric index as stored in the file
+    # example: rec[7]
     def [] index
+      return "#{@fields[index]}" if index.is_a? Integer
       "#{@columns[index.to_sym]}"
     end
 
+    # allow for column access via column name as a method
+    # example: rec.col_name
     def method_missing(id, *args)
       return @columns[id] if @columns[id]
       raise NoMethodError

data/lib/hivemeta/table.rb

@@ -11,7 +11,7 @@ module HiveMeta
       @path = nil
       @columns   = []
       @comments  = []
-      @delimiter = "\t"
+      @delimiter = "\001"
     end
 
     def to_s

metadata

@@ -5,8 +5,8 @@ version: !ruby/object:Gem::Version
   segments: 
   - 0
   - 0
-  - 2
-  version: 0.0.2
+  - 3
+  version: 0.0.3
 platform: ruby
 authors: 
 - Frank Fejes
@@ -14,7 +14,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2011-05-04 00:00:00 -05:00
+date: 2011-05-17 00:00:00 -05:00
 default_executable: 
 dependencies: []