Class: Embulk::Output::Vertica::OutputThread

Inherits:
Object
  • Object
show all
Defined in:
lib/embulk/output/vertica/output_thread.rb

Instance Method Summary collapse

Constructor Details

#initialize(task, schema, converters) ⇒ OutputThread

Returns a new instance of OutputThread.



32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
# File 'lib/embulk/output/vertica/output_thread.rb', line 32

def initialize(task, schema, converters)
  @task = task
  @schema = schema
  @queue = SizedQueue.new(1)
  @converters = converters
  @num_input_rows = 0
  @num_output_rows = 0
  @num_rejected_rows = 0
  @outer_thread = Thread.current
  @thread_active = false
  @progress_log_timer = Time.now

  case task['compress']
  when 'GZIP'
    @write_proc = self.method(:write_gzip)
  else
    @write_proc = self.method(:write_uncompressed)
  end
end

Instance Method Details

#abort_on_errorObject



199
200
201
# File 'lib/embulk/output/vertica/output_thread.rb', line 199

def abort_on_error
  @task['abort_on_error'] ? ' ABORT ON ERROR' : ''
end

#commitObject



144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
# File 'lib/embulk/output/vertica/output_thread.rb', line 144

def commit
  @thread_active = false
  if @thread.alive?
    @queue.push('finish')
    Embulk.logger.trace { "embulk-output-vertica: pushed finish" }
    Thread.pass
    @thread.join
  else
    raise RuntimeError, "embulk-output-vertica: thread died accidently"
  end

  task_report = {
    'num_input_rows' => @num_input_rows,
    'num_output_rows' => @num_output_rows,
    'num_rejected_rows' => @num_rejected_rows,
  }
end

#compressObject



191
192
193
# File 'lib/embulk/output/vertica/output_thread.rb', line 191

def compress
  " #{@task['compress']}"
end

#copy(conn, sql, &block) ⇒ Object

private



164
165
166
167
# File 'lib/embulk/output/vertica/output_thread.rb', line 164

def copy(conn, sql, &block)
  Embulk.logger.debug "embulk-output-vertica: #{sql}"
  results, rejects = conn.copy(sql, &block)
end

#copy_modeObject



195
196
197
# File 'lib/embulk/output/vertica/output_thread.rb', line 195

def copy_mode
  " #{@task['copy_mode']}"
end

#copy_sqlObject



169
170
171
# File 'lib/embulk/output/vertica/output_thread.rb', line 169

def copy_sql
  @copy_sql ||= "COPY #{quoted_schema}.#{quoted_temp_table} FROM STDIN#{compress}#{fjsonparser}#{copy_mode}#{abort_on_error} NO COMMIT"
end

#enqueue(page) ⇒ Object



52
53
54
55
56
57
58
59
60
# File 'lib/embulk/output/vertica/output_thread.rb', line 52

def enqueue(page)
  if @thread_active and @thread.alive?
    Embulk.logger.trace { "embulk-output-vertica: enqueue" }
    @queue.push(page)
  else
    Embulk.logger.info { "embulk-output-vertica: thread is dead, but still trying to enqueue" }
    raise RuntimeError, "embulk-output-vertica: thread is died, but still trying to enqueue"
  end
end

#fjsonparserObject



203
204
205
# File 'lib/embulk/output/vertica/output_thread.rb', line 203

def fjsonparser
  " PARSER fjsonparser(#{reject_on_materialized_type_error})"
end

#quoted_schemaObject



179
180
181
# File 'lib/embulk/output/vertica/output_thread.rb', line 179

def quoted_schema
  ::Jvertica.quote_identifier(@task['schema'])
end

#quoted_tableObject



183
184
185
# File 'lib/embulk/output/vertica/output_thread.rb', line 183

def quoted_table
  ::Jvertica.quote_identifier(@task['table'])
end

#quoted_temp_tableObject



187
188
189
# File 'lib/embulk/output/vertica/output_thread.rb', line 187

def quoted_temp_table
  ::Jvertica.quote_identifier(@task['temp_table'])
end

#reject_on_materialized_type_errorObject



207
208
209
# File 'lib/embulk/output/vertica/output_thread.rb', line 207

def reject_on_materialized_type_error
  @task['reject_on_materialized_type_error'] ? 'reject_on_materialized_type_error=true' : ''
end

#runObject



90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
# File 'lib/embulk/output/vertica/output_thread.rb', line 90

def run
  Embulk.logger.debug { "embulk-output-vertica: thread started" }
  Vertica.connect(@task) do |jv|
    json = nil # for log
    begin
      last_record = nil
      num_output_rows, rejects = copy(jv, copy_sql) do |stdin|
        while page = @queue.pop
          if page == 'finish'
            Embulk.logger.trace { "embulk-output-vertica: popped finish" }
            break
          end
          Embulk.logger.trace { "embulk-output-vertica: dequeued" }

          @write_proc.call(stdin, page) do |record|
            last_record = record
          end
        end
      end
      Embulk.logger.debug { "embulk-output-vertica: thread finished" }
      num_rejected_rows = rejects.size
      @num_output_rows += num_output_rows
      @num_rejected_rows += num_rejected_rows
      Embulk.logger.info { "embulk-output-vertica: COMMIT!" }
      jv.commit
      Embulk.logger.debug { "embulk-output-vertica: COMMITTED!" }
    rescue java.sql.SQLDataException => e
      if @task['reject_on_materialized_type_error'] and e.message =~ /Rejected by user-defined parser/
        Embulk.logger.warn "embulk-output-vertica: ROLLBACK! some of column types and values types do not fit #{json}"
      else
        Embulk.logger.warn "embulk-output-vertica: ROLLBACK!"
      end
      Embulk.logger.info { "embulk-output-vertica: last_record: #{to_json(last_record)}" }
      jv.rollback
      raise e # die transaction
    rescue => e
      Embulk.logger.warn "embulk-output-vertica: ROLLBACK!"
      jv.rollback
      raise e
    end
  end
rescue => e
  @thread_active = false # not to be enqueued any more
  while @queue.size > 0
    @queue.pop # dequeue all because some might be still trying @queue.push and get blocked, need to release
  end
  @outer_thread.raise e.class.new("#{e.message}\n  #{e.backtrace.join("\n  ")}")
end

#startObject



139
140
141
142
# File 'lib/embulk/output/vertica/output_thread.rb', line 139

def start
  @thread = Thread.new(&method(:run))
  @thread_active = true
end

#to_json(record) ⇒ Object



173
174
175
176
177
# File 'lib/embulk/output/vertica/output_thread.rb', line 173

def to_json(record)
  Hash[*(@schema.names.zip(record).map do |column_name, value|
    [column_name, @converters[column_name].call(value)]
  end.flatten!(1))].to_json
end

#write_buf(buf, page, &block) ⇒ Object



74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
# File 'lib/embulk/output/vertica/output_thread.rb', line 74

def write_buf(buf, page, &block)
  page.each do |record|
    yield(record) if block_given?
    Embulk.logger.trace { "embulk-output-vertica: record #{record}" }
    json = to_json(record)
    Embulk.logger.trace { "embulk-output-vertica: to_json #{json}" }
    buf << json << "\n"
    @num_input_rows += 1
  end
  now = Time.now
  if @progress_log_timer < now - 10 # once in 10 seconds
    @progress_log_timer = now
    Embulk.logger.info { "embulk-output-vertica: num_input_rows #{@num_input_rows}" }
  end
end

#write_gzip(io, page, &block) ⇒ Object



62
63
64
65
66
# File 'lib/embulk/output/vertica/output_thread.rb', line 62

def write_gzip(io, page, &block)
  buf = Zlib::Deflate.new
  write_buf(buf, page, &block)
  io << buf.finish
end

#write_uncompressed(io, page, &block) ⇒ Object



68
69
70
71
72
# File 'lib/embulk/output/vertica/output_thread.rb', line 68

def write_uncompressed(io, page, &block)
  buf = ''
  write_buf(buf, page, &block)
  io << buf
end