postgres_upsert 5.0.0 → 5.1.0
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +5 -13
- data/.rubocop.yml +57 -0
- data/.ruby-gemset +1 -0
- data/.ruby-version +1 -0
- data/.travis.yml +18 -3
- data/Gemfile +1 -2
- data/Gemfile.lock +171 -109
- data/README.md +8 -5
- data/app/assets/config/manifest.js +0 -0
- data/bin/setup +27 -0
- data/config/application.rb +0 -3
- data/config/database.yml +2 -0
- data/config/database.yml.travis +23 -0
- data/db/migrate/20150214192135_create_test_tables.rb +6 -1
- data/db/migrate/20150710162236_create_composite_models_table.rb +1 -1
- data/db/schema.rb +22 -17
- data/lib/postgres_upsert.rb +35 -6
- data/lib/postgres_upsert/model_to_model_adapter.rb +37 -0
- data/lib/postgres_upsert/read_adapters/active_record_adapter.rb +37 -0
- data/lib/postgres_upsert/read_adapters/file_adapter.rb +42 -0
- data/lib/postgres_upsert/read_adapters/io_adapter.rb +42 -0
- data/lib/postgres_upsert/result.rb +3 -4
- data/lib/postgres_upsert/table_writer.rb +7 -9
- data/lib/postgres_upsert/write_adapters/active_record_adapter.rb +36 -0
- data/lib/postgres_upsert/write_adapters/table_adapter.rb +56 -0
- data/lib/postgres_upsert/writer.rb +87 -69
- data/postgres_upsert.gemspec +6 -3
- data/spec/composite_key_spec.rb +0 -6
- data/spec/fixtures/comma_with_header_duplicate.csv +3 -0
- data/spec/fixtures/test_model_copy.rb +4 -0
- data/spec/from_table_spec.rb +40 -0
- data/spec/pg_upsert_csv_spec.rb +11 -10
- data/spec/rails_helper.rb +1 -0
- data/spec/spec_helper.rb +5 -2
- metadata +77 -31
- data/VERSION +0 -1
File without changes
|
data/bin/setup
CHANGED
@@ -4,6 +4,19 @@ require 'pathname'
|
|
4
4
|
# path to your application root.
|
5
5
|
APP_ROOT = Pathname.new File.expand_path('../../', __FILE__)
|
6
6
|
|
7
|
+
|
8
|
+
def silently command
|
9
|
+
system command + " > /dev/null 2>&1"
|
10
|
+
end
|
11
|
+
|
12
|
+
def successfully command
|
13
|
+
silently command or fail "Error in script: " + command
|
14
|
+
end
|
15
|
+
|
16
|
+
def announce message
|
17
|
+
puts "\n== " + message + " =="
|
18
|
+
end
|
19
|
+
|
7
20
|
Dir.chdir APP_ROOT do
|
8
21
|
# This script is a starting point to setup your application.
|
9
22
|
# Add necessary setup steps to this file:
|
@@ -17,6 +30,20 @@ Dir.chdir APP_ROOT do
|
|
17
30
|
# system "cp config/database.yml.sample config/database.yml"
|
18
31
|
# end
|
19
32
|
|
33
|
+
unless silently "ls -A /usr/local/var/postgres"
|
34
|
+
announce "initializing postgres db in /usr/local/var/postgres"
|
35
|
+
silently "initdb /usr/local/var/postgres"
|
36
|
+
|
37
|
+
announce "creating superuser 'postgres'. Hope that's cool?"
|
38
|
+
silently "createuser -s postgres"
|
39
|
+
end
|
40
|
+
|
41
|
+
unless silently "ls -A ~/Library/LaunchAgents/homebrew.mxcl.postgresql.plist"
|
42
|
+
announce "setting up Postgres to start on launch"
|
43
|
+
silently "ln -sfv /usr/local/opt/postgresql/*.plist ~/Library/LaunchAgents"
|
44
|
+
silently "launchctl load -w ~/Library/LaunchAgents/homebrew.mxcl.postgresql.plist"
|
45
|
+
end
|
46
|
+
|
20
47
|
puts "\n== Preparing database =="
|
21
48
|
system "bin/rake db:setup"
|
22
49
|
|
data/config/application.rb
CHANGED
@@ -18,7 +18,4 @@ class Application < Rails::Application
|
|
18
18
|
# The default locale is :en and all translations from config/locales/*.rb,yml are auto loaded.
|
19
19
|
# config.i18n.load_path += Dir[Rails.root.join('my', 'locales', '*.{rb,yml}').to_s]
|
20
20
|
# config.i18n.default_locale = :de
|
21
|
-
|
22
|
-
# Do not swallow errors in after_commit/after_rollback callbacks.
|
23
|
-
config.active_record.raise_in_transactional_callbacks = true
|
24
21
|
end
|
data/config/database.yml
CHANGED
@@ -0,0 +1,23 @@
|
|
1
|
+
# SQLite version 3.x
|
2
|
+
# gem 'activerecord-jdbcsqlite3-adapter'
|
3
|
+
#
|
4
|
+
# Configure Using Gemfile
|
5
|
+
# gem 'activerecord-jdbcsqlite3-adapter'
|
6
|
+
#
|
7
|
+
default: &default
|
8
|
+
adapter: postgresql
|
9
|
+
host: localhost
|
10
|
+
port: 5432
|
11
|
+
pool: 5
|
12
|
+
username: postgres
|
13
|
+
|
14
|
+
development:
|
15
|
+
<<: *default
|
16
|
+
database: ar_pg_copy_dev
|
17
|
+
|
18
|
+
# Warning: The database defined as "test" will be erased and
|
19
|
+
# re-generated from your development database when you run "rake".
|
20
|
+
# Do not set this db to the same as development or production.
|
21
|
+
test:
|
22
|
+
<<: *default
|
23
|
+
database: ar_pg_copy_test
|
@@ -1,10 +1,15 @@
|
|
1
|
-
class CreateTestTables < ActiveRecord::Migration
|
1
|
+
class CreateTestTables < ActiveRecord::Migration[6.0]
|
2
2
|
def change
|
3
3
|
create_table :test_models do |t|
|
4
4
|
t.string :data
|
5
5
|
t.timestamps
|
6
6
|
end
|
7
7
|
|
8
|
+
create_table :test_model_copies do |t|
|
9
|
+
t.string :data
|
10
|
+
t.timestamps
|
11
|
+
end
|
12
|
+
|
8
13
|
create_table :three_columns do |t|
|
9
14
|
t.string :data
|
10
15
|
t.string :extra
|
data/db/schema.rb
CHANGED
@@ -1,17 +1,16 @@
|
|
1
|
-
# encoding: UTF-8
|
2
1
|
# This file is auto-generated from the current state of the database. Instead
|
3
2
|
# of editing this file, please use the migrations feature of Active Record to
|
4
3
|
# incrementally modify your database, and then regenerate this schema definition.
|
5
4
|
#
|
6
|
-
#
|
7
|
-
#
|
8
|
-
#
|
9
|
-
# from scratch.
|
10
|
-
#
|
5
|
+
# This file is the source Rails uses to define your schema when running `rails
|
6
|
+
# db:schema:load`. When creating a new database, `rails db:schema:load` tends to
|
7
|
+
# be faster and is potentially less error prone than running all of your
|
8
|
+
# migrations from scratch. Old migrations may fail to apply correctly if those
|
9
|
+
# migrations use external dependencies or application code.
|
11
10
|
#
|
12
11
|
# It's strongly recommended that you check this file into your version control system.
|
13
12
|
|
14
|
-
ActiveRecord::Schema.define(version:
|
13
|
+
ActiveRecord::Schema.define(version: 2015_07_10_162236) do
|
15
14
|
|
16
15
|
# These are extensions that must be enabled in order to support this database
|
17
16
|
enable_extension "plpgsql"
|
@@ -19,25 +18,31 @@ ActiveRecord::Schema.define(version: 20150710162236) do
|
|
19
18
|
create_table "composite_key_models", force: :cascade do |t|
|
20
19
|
t.integer "comp_key_1"
|
21
20
|
t.integer "comp_key_2"
|
22
|
-
t.string
|
21
|
+
t.string "data"
|
23
22
|
end
|
24
23
|
|
25
24
|
create_table "reserved_word_models", force: :cascade do |t|
|
26
|
-
t.string "select"
|
27
|
-
t.string "group"
|
25
|
+
t.string "select"
|
26
|
+
t.string "group"
|
27
|
+
end
|
28
|
+
|
29
|
+
create_table "test_model_copies", force: :cascade do |t|
|
30
|
+
t.string "data"
|
31
|
+
t.datetime "created_at", precision: 6, null: false
|
32
|
+
t.datetime "updated_at", precision: 6, null: false
|
28
33
|
end
|
29
34
|
|
30
35
|
create_table "test_models", force: :cascade do |t|
|
31
|
-
t.string
|
32
|
-
t.datetime "created_at"
|
33
|
-
t.datetime "updated_at"
|
36
|
+
t.string "data"
|
37
|
+
t.datetime "created_at", precision: 6, null: false
|
38
|
+
t.datetime "updated_at", precision: 6, null: false
|
34
39
|
end
|
35
40
|
|
36
41
|
create_table "three_columns", force: :cascade do |t|
|
37
|
-
t.string
|
38
|
-
t.string
|
39
|
-
t.datetime "created_at"
|
40
|
-
t.datetime "updated_at"
|
42
|
+
t.string "data"
|
43
|
+
t.string "extra"
|
44
|
+
t.datetime "created_at", precision: 6, null: false
|
45
|
+
t.datetime "updated_at", precision: 6, null: false
|
41
46
|
end
|
42
47
|
|
43
48
|
end
|
data/lib/postgres_upsert.rb
CHANGED
@@ -1,17 +1,46 @@
|
|
1
1
|
require 'rubygems'
|
2
2
|
require 'active_record'
|
3
|
+
require 'postgres_upsert/read_adapters/active_record_adapter'
|
4
|
+
require 'postgres_upsert/read_adapters/file_adapter'
|
5
|
+
require 'postgres_upsert/read_adapters/io_adapter'
|
6
|
+
require 'postgres_upsert/write_adapters/active_record_adapter'
|
7
|
+
require 'postgres_upsert/write_adapters/table_adapter'
|
3
8
|
require 'postgres_upsert/writer'
|
4
9
|
require 'postgres_upsert/table_writer'
|
10
|
+
require 'postgres_upsert/model_to_model_adapter'
|
5
11
|
require 'postgres_upsert/result'
|
6
12
|
require 'rails'
|
7
13
|
|
8
14
|
module PostgresUpsert
|
9
|
-
|
10
15
|
class << self
|
11
|
-
def write
|
12
|
-
|
13
|
-
|
14
|
-
|
16
|
+
def write(destination, source, options = {})
|
17
|
+
read_adapter = read_adapter(source).new(source, options)
|
18
|
+
write_adapter = write_adapter(destination).new(destination, options)
|
19
|
+
Writer.new(destination, write_adapter, read_adapter, options).write
|
20
|
+
end
|
21
|
+
|
22
|
+
def read_adapter(source)
|
23
|
+
if [StringIO, File].include?(source.class)
|
24
|
+
ReadAdapters::IOAdapter
|
25
|
+
elsif [String].include?(source.class)
|
26
|
+
ReadAdapters::FileAdapter
|
27
|
+
elsif source < ActiveRecord::Base
|
28
|
+
ReadAdapters::ActiveRecordAdapter
|
29
|
+
else
|
30
|
+
raise "Source must be a Filename string, StringIO of data, or a ActiveRecord Class."
|
31
|
+
end
|
32
|
+
end
|
33
|
+
|
34
|
+
def write_adapter(destination)
|
35
|
+
if [String].include?(destination.class)
|
36
|
+
WriteAdapters::TableAdapter
|
37
|
+
elsif destination <= ActiveRecord::Base
|
38
|
+
WriteAdapters::ActiveRecordAdapter
|
39
|
+
# elsif source < ActiveRecord::Base && destination < ActiveRecord::Base
|
40
|
+
#ModelToModelAdapter
|
41
|
+
else
|
42
|
+
raise "Destination must be an ActiveRecord class or a table name string"
|
43
|
+
end
|
15
44
|
end
|
16
45
|
end
|
17
|
-
end
|
46
|
+
end
|
@@ -0,0 +1,37 @@
|
|
1
|
+
require 'csv'
|
2
|
+
|
3
|
+
module PostgresUpsert
|
4
|
+
class ModelToModelAdapter
|
5
|
+
def initialize(destination_model, source_model, options = {})
|
6
|
+
@destination_model = destination_model
|
7
|
+
@source_model = source_model
|
8
|
+
@options = options
|
9
|
+
end
|
10
|
+
|
11
|
+
def write
|
12
|
+
source_table = @source_model.table_name
|
13
|
+
source_conn = @source_model.connection.raw_connection
|
14
|
+
|
15
|
+
to_stdout_sql = "COPY #{source_table} TO STDOUT"
|
16
|
+
|
17
|
+
csv_string = CSV.generate do |csv|
|
18
|
+
csv << @source_model.column_names # CSV header row
|
19
|
+
source_conn.copy_data(to_stdout_sql) do
|
20
|
+
while (line = source_conn.get_copy_data) do
|
21
|
+
csv << line.split("\t")
|
22
|
+
end
|
23
|
+
end
|
24
|
+
end
|
25
|
+
io = StringIO.new(csv_string)
|
26
|
+
Writer.new(@destination_model, io, @options).write
|
27
|
+
end
|
28
|
+
|
29
|
+
private
|
30
|
+
|
31
|
+
def get_columns
|
32
|
+
# columns_list = @options[:columns]
|
33
|
+
# columns_list ||=
|
34
|
+
@source.column_names
|
35
|
+
end
|
36
|
+
end
|
37
|
+
end
|
@@ -0,0 +1,37 @@
|
|
1
|
+
module PostgresUpsert
|
2
|
+
module ReadAdapters
|
3
|
+
class ActiveRecordAdapter
|
4
|
+
def initialize(source, options)
|
5
|
+
@options = sanitize_options(options)
|
6
|
+
@source = source
|
7
|
+
end
|
8
|
+
|
9
|
+
def sanitize_options(options)
|
10
|
+
options.slice(
|
11
|
+
:columns, :map, :unique_key
|
12
|
+
)
|
13
|
+
end
|
14
|
+
|
15
|
+
def continuous_write_enabled
|
16
|
+
false
|
17
|
+
end
|
18
|
+
|
19
|
+
def gets(&block)
|
20
|
+
batch_size = 1_000
|
21
|
+
line = ""
|
22
|
+
conn = @source.connection.raw_connection
|
23
|
+
|
24
|
+
conn.copy_data("COPY #{@source.table_name} TO STDOUT") do
|
25
|
+
while (line_read = conn.get_copy_data) do
|
26
|
+
line << line_read
|
27
|
+
end
|
28
|
+
end
|
29
|
+
yield line
|
30
|
+
end
|
31
|
+
|
32
|
+
def columns
|
33
|
+
@source.column_names
|
34
|
+
end
|
35
|
+
end
|
36
|
+
end
|
37
|
+
end
|
@@ -0,0 +1,42 @@
|
|
1
|
+
module PostgresUpsert
|
2
|
+
module ReadAdapters
|
3
|
+
class FileAdapter
|
4
|
+
def initialize(source, options)
|
5
|
+
@options = sanitize_options(options)
|
6
|
+
@source = File.open(source, 'r')
|
7
|
+
end
|
8
|
+
|
9
|
+
def sanitize_options(options)
|
10
|
+
options.slice(
|
11
|
+
:delimiter, :header, :columns, :map, :unique_key
|
12
|
+
).reverse_merge(
|
13
|
+
header: true,
|
14
|
+
delimiter: ',',
|
15
|
+
)
|
16
|
+
end
|
17
|
+
|
18
|
+
def continuous_write_enabled
|
19
|
+
true
|
20
|
+
end
|
21
|
+
|
22
|
+
def gets
|
23
|
+
@source.gets
|
24
|
+
end
|
25
|
+
|
26
|
+
def columns
|
27
|
+
@columns ||= begin
|
28
|
+
columns_list = @options[:columns] ? @options[:columns].map(&:to_s) : []
|
29
|
+
if @options[:header]
|
30
|
+
# if header is present, we need to strip it from io, whether we use it for the columns list or not.
|
31
|
+
line = gets
|
32
|
+
if columns_list.empty?
|
33
|
+
columns_list = line.strip.split(@options[:delimiter])
|
34
|
+
end
|
35
|
+
end
|
36
|
+
columns_list = columns_list.map { |c| @options[:map][c.to_s] } if @options[:map]
|
37
|
+
columns_list
|
38
|
+
end
|
39
|
+
end
|
40
|
+
end
|
41
|
+
end
|
42
|
+
end
|
@@ -0,0 +1,42 @@
|
|
1
|
+
module PostgresUpsert
|
2
|
+
module ReadAdapters
|
3
|
+
class IOAdapter
|
4
|
+
def initialize(source, options)
|
5
|
+
@options = sanitize_options(options)
|
6
|
+
@source = source
|
7
|
+
end
|
8
|
+
|
9
|
+
def sanitize_options(options)
|
10
|
+
options.slice(
|
11
|
+
:delimiter, :header, :columns, :map, :unique_key
|
12
|
+
).reverse_merge(
|
13
|
+
header: true,
|
14
|
+
delimiter: ',',
|
15
|
+
)
|
16
|
+
end
|
17
|
+
|
18
|
+
def continuous_write_enabled
|
19
|
+
true
|
20
|
+
end
|
21
|
+
|
22
|
+
def gets
|
23
|
+
@source.gets
|
24
|
+
end
|
25
|
+
|
26
|
+
def columns
|
27
|
+
@columns ||= begin
|
28
|
+
columns_list = @options[:columns] ? @options[:columns].map(&:to_s) : []
|
29
|
+
if @options[:header]
|
30
|
+
# if header is present, we need to strip it from io, whether we use it for the columns list or not.
|
31
|
+
line = gets
|
32
|
+
if columns_list.empty?
|
33
|
+
columns_list = line.strip.split(@options[:delimiter])
|
34
|
+
end
|
35
|
+
end
|
36
|
+
columns_list = columns_list.map { |c| @options[:map][c.to_s] } if @options[:map]
|
37
|
+
columns_list
|
38
|
+
end
|
39
|
+
end
|
40
|
+
end
|
41
|
+
end
|
42
|
+
end
|
@@ -4,8 +4,8 @@ module PostgresUpsert
|
|
4
4
|
|
5
5
|
def initialize(insert_result, update_result, copy_result)
|
6
6
|
@inserted = insert_result ? insert_result.cmd_tuples : 0
|
7
|
-
@updated = update_result ? update_result.cmd_tuples : 0
|
8
|
-
@copied = copy_result ? copy_result.cmd_tuples : 0
|
7
|
+
@updated = update_result ? update_result.cmd_tuples : 0
|
8
|
+
@copied = copy_result ? copy_result.cmd_tuples : 0
|
9
9
|
end
|
10
10
|
|
11
11
|
def changed_rows
|
@@ -20,5 +20,4 @@ module PostgresUpsert
|
|
20
20
|
@updated
|
21
21
|
end
|
22
22
|
end
|
23
|
-
end
|
24
|
-
|
23
|
+
end
|
@@ -1,8 +1,7 @@
|
|
1
1
|
module PostgresUpsert
|
2
|
-
# alternate version of PostgresUpsert::Writer which does not rely on AR table information.
|
3
|
-
#
|
2
|
+
# alternate version of PostgresUpsert::Writer which does not rely on AR table information.
|
3
|
+
# We can use this model to upsert data into views, or tables not associated to rails models
|
4
4
|
class TableWriter < Writer
|
5
|
-
|
6
5
|
def initialize(table_name, source, options = {})
|
7
6
|
@table_name = table_name
|
8
7
|
super(nil, source, options)
|
@@ -16,7 +15,7 @@ module PostgresUpsert
|
|
16
15
|
|
17
16
|
def primary_key
|
18
17
|
@primary_key ||= begin
|
19
|
-
query = <<-
|
18
|
+
query = <<-SELECT_KEY
|
20
19
|
SELECT
|
21
20
|
pg_attribute.attname,
|
22
21
|
format_type(pg_attribute.atttypid, pg_attribute.atttypmod)
|
@@ -27,24 +26,23 @@ module PostgresUpsert
|
|
27
26
|
pg_attribute.attrelid = pg_class.oid AND
|
28
27
|
pg_attribute.attnum = any(pg_index.indkey)
|
29
28
|
AND indisprimary
|
30
|
-
|
29
|
+
SELECT_KEY
|
31
30
|
|
32
31
|
pg_result = ActiveRecord::Base.connection.execute query
|
33
|
-
pg_result.each{ |row| return row['attname'] }
|
32
|
+
pg_result.each { |row| return row['attname'] }
|
34
33
|
end
|
35
34
|
end
|
36
35
|
|
37
|
-
def
|
36
|
+
def destination_columns
|
38
37
|
@column_names ||= begin
|
39
38
|
query = "SELECT * FROM information_schema.columns WHERE TABLE_NAME = '#{@table_name}'"
|
40
39
|
pg_result = ActiveRecord::Base.connection.execute query
|
41
|
-
pg_result.map{ |row| row['column_name'] }
|
40
|
+
pg_result.map { |row| row['column_name'] }
|
42
41
|
end
|
43
42
|
end
|
44
43
|
|
45
44
|
def quoted_table_name
|
46
45
|
@quoted_table_name ||= ActiveRecord::Base.connection.quote_table_name(@table_name)
|
47
46
|
end
|
48
|
-
|
49
47
|
end
|
50
48
|
end
|