Class: DbAgile::Command::Bulk::Import

Inherits:
DbAgile::Command show all
Includes:
Commons, Tools::Tuple
Defined in:
lib/dbagile/command/bulk/import.rb

Overview

Import a table from (csv, json, yaml, ruby)

Usage: dba #DbAgile::Command#command_name [OPTIONS] TABLE [FILE]

Constant Summary

Constants inherited from DbAgile::Command

CATEGORIES, CATEGORY_NAMES

Instance Attribute Summary collapse

Attributes included from Commons

#allbut, #conn_options, #dataset, #format, #io_options, #select, #type_system

Attributes inherited from DbAgile::Command

#environment

Attributes included from ClassMethods

#description, #summary, #usage

Instance Method Summary collapse

Methods included from Tools::Tuple

#check_tuple_heading, #tuple_has_key?, #tuple_heading, #tuple_key, #tuple_project, #tuple_to_querystring

Methods included from Commons

#add_csv_options, #add_html_output_options, #add_input_format_options, #add_json_output_options, #add_output_format_options, #add_select_options, #add_text_output_options, #add_typesafe_options, #set_default_options

Methods inherited from DbAgile::Command

#category, #check_command, #command_name, #description, #initialize, #options, #run, #set_default_options, #show_help, #summary, #unsecure_run, #usage

Methods included from ClassMethods

#build_command_options, #build_me, #category, #command_for, #command_name, #command_name_of, #each_subclass, #inherited, #ruby_method_for, #subclasses

Methods included from Robust

#ambigous_argument_list!, #assumption_error!, #bad_argument_list!, #has_command!, #is_in!, #valid_argument_list!, #valid_read_file!

Methods included from DbAgile::Core::IO::Robustness

#has_database!, #valid_database_name!, #valid_database_uri!, #valid_schema_files!

Constructor Details

This class inherits a constructor from DbAgile::Command

Instance Attribute Details

#create_table ⇒ Object

Create table?



21
22
23
 
# File 'lib/dbagile/command/bulk/import.rb', line 21

def create_table
  @create_table
end

#drop_table ⇒ Object

Drop table?



18
19
20
 
# File 'lib/dbagile/command/bulk/import.rb', line 18

def drop_table
  @drop_table
end

#input_file ⇒ Object

Input file to use



15
16
17
 
# File 'lib/dbagile/command/bulk/import.rb', line 15

def input_file
  @input_file
end

#truncate_table ⇒ Object

Truncate table?



24
25
26
 
# File 'lib/dbagile/command/bulk/import.rb', line 24

def truncate_table
  @truncate_table
end

#update_on_existing ⇒ Object

Update tuple when key already exists?



27
28
29
 
# File 'lib/dbagile/command/bulk/import.rb', line 27

def update_on_existing
  @update_on_existing
end

Instance Method Details

#add_options(opt) ⇒ Object

Contribute to options



30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
 
# File 'lib/dbagile/command/bulk/import.rb', line 30

def add_options(opt)
  # Main output options
  opt.separator "\nInput options:"
  opt.on("--input=FILE", "-i", "Read input from FILE (stdin by default)") do |value|
    self.input_file = value
  end

  opt.separator "\nSQL options:"
  opt.on("--create-table", "-c", "Create table before inserting values") do |value|
    self.drop_table = true
    self.create_table = true
    self.truncate_table = false
  end
  opt.on("--drop-table", "-d", "Drop table before inserting values") do |value|
    self.drop_table = true
    self.create_table = true
    self.truncate_table = false
  end
  opt.on("--truncate", "-t", "Truncate table before inserting values") do |value|
    self.truncate_table = true
  end
  opt.on('--trace-sql', "Trace SQL statements on STDOUT (but executes them)") do |value|
    self.conn_options[:trace_sql] = true
  end
  opt.on("--update", "-u", "Update tuple instead of insert when key already exists") do |value|
    self.update_on_existing = true
  end
  opt.on('--dry-run', "Trace SQL statements on STDOUT only, do nothing on the database") do |value|
    self.conn_options[:trace_sql] = true
    self.conn_options[:trace_only] = true
    self.conn_options[:trace_buffer] = environment.output_buffer
  end

  opt.separator "\nRecognized format options:"
  add_input_format_options(opt)
  add_typesafe_options(opt)

  # CSV output options
  opt.separator "\nCSV options:"
  add_csv_input_options(opt)
end

#execute_command ⇒ Object

Executes the command



114
115
116
117
118
119
120
121
122
123
124
125
126
127
 
# File 'lib/dbagile/command/bulk/import.rb', line 114

def execute_command
  with_current_connection(conn_options) do |connection|
          
    # Make the job now
    connection.transaction do |t|
      first = true
      with_emitter do |tuple|
        make_the_job(t, tuple, first)
        first = false
      end
    end
  
  end
end

#find_key(t, tuple) ⇒ Object 



129
130
131
132
133
 
# File 'lib/dbagile/command/bulk/import.rb', line 129

def find_key(t, tuple)
  t.keys(self.dataset).find{|key|
    key.all?{|k| tuple.has_key?(k)}
  }
end

#handle_schema_modification(t, tuple) ⇒ Object

Handles the schema modifications



159
160
161
162
163
164
165
166
167
168
169
170
171
172
 
# File 'lib/dbagile/command/bulk/import.rb', line 159

def handle_schema_modification(t, tuple)
  # Handle table creation/deletion/truncation
  if drop_table
    t.drop_table(self.dataset) if t.has_table?(self.dataset)
  end
  if create_table
    heading = tuple_heading(tuple)
    heading = check_tuple_heading(heading, environment)
    t.create_table(self.dataset, heading)
  end
  if truncate_table
    t.delete(self.dataset, {})
  end
end

#make_the_job(t, tuple, first = true) ⇒ Object

Makes the insert job



136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
 
# File 'lib/dbagile/command/bulk/import.rb', line 136

def make_the_job(t, tuple, first = true)
  handle_schema_modification(t, tuple) if first
  if self.update_on_existing 
    
    # find a candidate key
    key = find_key(t, tuple)
    if key.nil?
      raise DbAgile::CandidateKeyNotFoundError, "Unable to find a candidate key for UPDATE"
    end

    # Make insert or update according to key
    key_value = tuple_project(tuple, key)
    if t.exists?(self.dataset, key_value)
      t.update(self.dataset, tuple, key_value)
    else
      t.insert(self.dataset, tuple)
    end
  else
    t.insert(self.dataset, tuple)
  end
end

#normalize_pending_arguments(arguments) ⇒ Object

Normalizes the pending arguments



73
74
75
76
77
78
79
80
81
82
83
 
# File 'lib/dbagile/command/bulk/import.rb', line 73

def normalize_pending_arguments(arguments)
  case arguments.size
    when 1
      self.dataset = valid_argument_list!(arguments, Symbol)
    when 2
      raise OptionParser::AmbiguousArgument, '--input-file=#{self.input_file}' if self.input_file
      self.dataset, self.input_file = valid_argument_list!(arguments, Symbol, String)
    else
      bad_argument_list!(arguments)
  end
end

#with_emitter(&block) ⇒ Object

Yields the block with each emitted tuple



95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
 
# File 'lib/dbagile/command/bulk/import.rb', line 95

def with_emitter(&block)
  with_io{|io|
    options = io_options[self.format]
    options[:type_system] = self.type_system if self.type_system
    options[:input_file] = self.input_file
    case self.format
      when :csv
        DbAgile::IO::CSV::from_csv(io, options, &block)
      when :json
        DbAgile::IO::JSON::from_json(io, options, &block)
      when :yaml
        DbAgile::IO::YAML::from_yaml(io, options, &block)
      when :ruby
        DbAgile::IO::Ruby::from_ruby(io, options, &block)
    end
  }
end

#with_io(&block) ⇒ Object

Yields the block with the IO object to use for input



86
87
88
89
90
91
92
 
# File 'lib/dbagile/command/bulk/import.rb', line 86

def with_io(&block)
  if self.input_file
    File.open(self.input_file, "r", &block)
  else
    block.call(environment.input_buffer)
  end
end