postgres_upsert 2.0.0 → 5.1.0
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +5 -5
- data/.gitignore +3 -0
- data/.rubocop.yml +57 -0
- data/.ruby-gemset +1 -0
- data/.ruby-version +1 -0
- data/.travis.yml +20 -0
- data/Gemfile +1 -2
- data/Gemfile.lock +175 -85
- data/README.md +117 -41
- data/Rakefile +4 -16
- data/app/assets/config/manifest.js +0 -0
- data/bin/bundle +3 -0
- data/bin/rails +4 -0
- data/bin/rake +4 -0
- data/bin/setup +56 -0
- data/config.ru +4 -0
- data/config/application.rb +21 -0
- data/config/boot.rb +3 -0
- data/config/database.yml +24 -0
- data/config/database.yml.travis +23 -0
- data/config/environment.rb +5 -0
- data/config/environments/development.rb +41 -0
- data/config/environments/production.rb +79 -0
- data/config/environments/test.rb +42 -0
- data/config/locales/en.yml +23 -0
- data/config/routes.rb +56 -0
- data/config/secrets.yml +22 -0
- data/db/migrate/20150214192135_create_test_tables.rb +24 -0
- data/db/migrate/20150710162236_create_composite_models_table.rb +9 -0
- data/db/schema.rb +48 -0
- data/db/seeds.rb +7 -0
- data/lib/postgres_upsert.rb +38 -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 +23 -0
- data/lib/postgres_upsert/table_writer.rb +48 -0
- 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 +130 -92
- data/postgres_upsert.gemspec +7 -4
- data/spec/composite_key_spec.rb +50 -0
- data/spec/fixtures/comma_with_header_duplicate.csv +3 -0
- data/spec/fixtures/composite_key_model.rb +4 -0
- data/spec/fixtures/composite_key_with_header.csv +3 -0
- data/spec/fixtures/composite_nonkey_with_header.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 +93 -35
- data/spec/rails_helper.rb +1 -0
- data/spec/spec_helper.rb +9 -37
- metadata +106 -37
- data/VERSION +0 -1
- data/lib/postgres_upsert/active_record.rb +0 -13
- data/spec/fixtures/2_col_binary_data.dat +0 -0
- data/spec/pg_upsert_binary_spec.rb +0 -35
- data/spec/spec.opts +0 -1
@@ -0,0 +1,24 @@
|
|
1
|
+
class CreateTestTables < ActiveRecord::Migration[6.0]
|
2
|
+
def change
|
3
|
+
create_table :test_models do |t|
|
4
|
+
t.string :data
|
5
|
+
t.timestamps
|
6
|
+
end
|
7
|
+
|
8
|
+
create_table :test_model_copies do |t|
|
9
|
+
t.string :data
|
10
|
+
t.timestamps
|
11
|
+
end
|
12
|
+
|
13
|
+
create_table :three_columns do |t|
|
14
|
+
t.string :data
|
15
|
+
t.string :extra
|
16
|
+
t.timestamps
|
17
|
+
end
|
18
|
+
|
19
|
+
create_table :reserved_word_models do |t|
|
20
|
+
t.string :select
|
21
|
+
t.string :group
|
22
|
+
end
|
23
|
+
end
|
24
|
+
end
|
data/db/schema.rb
ADDED
@@ -0,0 +1,48 @@
|
|
1
|
+
# This file is auto-generated from the current state of the database. Instead
|
2
|
+
# of editing this file, please use the migrations feature of Active Record to
|
3
|
+
# incrementally modify your database, and then regenerate this schema definition.
|
4
|
+
#
|
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.
|
10
|
+
#
|
11
|
+
# It's strongly recommended that you check this file into your version control system.
|
12
|
+
|
13
|
+
ActiveRecord::Schema.define(version: 2015_07_10_162236) do
|
14
|
+
|
15
|
+
# These are extensions that must be enabled in order to support this database
|
16
|
+
enable_extension "plpgsql"
|
17
|
+
|
18
|
+
create_table "composite_key_models", force: :cascade do |t|
|
19
|
+
t.integer "comp_key_1"
|
20
|
+
t.integer "comp_key_2"
|
21
|
+
t.string "data"
|
22
|
+
end
|
23
|
+
|
24
|
+
create_table "reserved_word_models", force: :cascade do |t|
|
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
|
33
|
+
end
|
34
|
+
|
35
|
+
create_table "test_models", force: :cascade do |t|
|
36
|
+
t.string "data"
|
37
|
+
t.datetime "created_at", precision: 6, null: false
|
38
|
+
t.datetime "updated_at", precision: 6, null: false
|
39
|
+
end
|
40
|
+
|
41
|
+
create_table "three_columns", force: :cascade do |t|
|
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
|
46
|
+
end
|
47
|
+
|
48
|
+
end
|
data/db/seeds.rb
ADDED
@@ -0,0 +1,7 @@
|
|
1
|
+
# This file should contain all the record creation needed to seed the database with its default values.
|
2
|
+
# The data can then be loaded with the rake db:seed (or created alongside the db with db:setup).
|
3
|
+
#
|
4
|
+
# Examples:
|
5
|
+
#
|
6
|
+
# cities = City.create([{ name: 'Chicago' }, { name: 'Copenhagen' }])
|
7
|
+
# Mayor.create(name: 'Emanuel', city: cities.first)
|
data/lib/postgres_upsert.rb
CHANGED
@@ -1,14 +1,46 @@
|
|
1
1
|
require 'rubygems'
|
2
2
|
require 'active_record'
|
3
|
-
require 'postgres_upsert/
|
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'
|
4
8
|
require 'postgres_upsert/writer'
|
9
|
+
require 'postgres_upsert/table_writer'
|
10
|
+
require 'postgres_upsert/model_to_model_adapter'
|
11
|
+
require 'postgres_upsert/result'
|
5
12
|
require 'rails'
|
6
13
|
|
7
|
-
|
14
|
+
module PostgresUpsert
|
15
|
+
class << self
|
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
|
8
33
|
|
9
|
-
|
10
|
-
|
11
|
-
|
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
|
12
44
|
end
|
13
45
|
end
|
14
|
-
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
|
@@ -0,0 +1,23 @@
|
|
1
|
+
module PostgresUpsert
|
2
|
+
class Result
|
3
|
+
attr_reader :inserted, :updated
|
4
|
+
|
5
|
+
def initialize(insert_result, update_result, copy_result)
|
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
|
9
|
+
end
|
10
|
+
|
11
|
+
def changed_rows
|
12
|
+
@inserted + @updated
|
13
|
+
end
|
14
|
+
|
15
|
+
def copied_rows
|
16
|
+
@copied
|
17
|
+
end
|
18
|
+
|
19
|
+
def updated_rows
|
20
|
+
@updated
|
21
|
+
end
|
22
|
+
end
|
23
|
+
end
|
@@ -0,0 +1,48 @@
|
|
1
|
+
module PostgresUpsert
|
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
|
+
class TableWriter < Writer
|
5
|
+
def initialize(table_name, source, options = {})
|
6
|
+
@table_name = table_name
|
7
|
+
super(nil, source, options)
|
8
|
+
end
|
9
|
+
|
10
|
+
private
|
11
|
+
|
12
|
+
def database_connection
|
13
|
+
ActiveRecord::Base.connection
|
14
|
+
end
|
15
|
+
|
16
|
+
def primary_key
|
17
|
+
@primary_key ||= begin
|
18
|
+
query = <<-SELECT_KEY
|
19
|
+
SELECT
|
20
|
+
pg_attribute.attname,
|
21
|
+
format_type(pg_attribute.atttypid, pg_attribute.atttypmod)
|
22
|
+
FROM pg_index, pg_class, pg_attribute
|
23
|
+
WHERE
|
24
|
+
pg_class.oid = '#{@table_name}'::regclass AND
|
25
|
+
indrelid = pg_class.oid AND
|
26
|
+
pg_attribute.attrelid = pg_class.oid AND
|
27
|
+
pg_attribute.attnum = any(pg_index.indkey)
|
28
|
+
AND indisprimary
|
29
|
+
SELECT_KEY
|
30
|
+
|
31
|
+
pg_result = ActiveRecord::Base.connection.execute query
|
32
|
+
pg_result.each { |row| return row['attname'] }
|
33
|
+
end
|
34
|
+
end
|
35
|
+
|
36
|
+
def destination_columns
|
37
|
+
@column_names ||= begin
|
38
|
+
query = "SELECT * FROM information_schema.columns WHERE TABLE_NAME = '#{@table_name}'"
|
39
|
+
pg_result = ActiveRecord::Base.connection.execute query
|
40
|
+
pg_result.map { |row| row['column_name'] }
|
41
|
+
end
|
42
|
+
end
|
43
|
+
|
44
|
+
def quoted_table_name
|
45
|
+
@quoted_table_name ||= ActiveRecord::Base.connection.quote_table_name(@table_name)
|
46
|
+
end
|
47
|
+
end
|
48
|
+
end
|
@@ -0,0 +1,36 @@
|
|
1
|
+
module PostgresUpsert
|
2
|
+
module WriteAdapters
|
3
|
+
class ActiveRecordAdapter
|
4
|
+
def initialize(destination, options)
|
5
|
+
@destination = destination
|
6
|
+
@options = sanitize_options(options)
|
7
|
+
|
8
|
+
end
|
9
|
+
|
10
|
+
def sanitize_options(options)
|
11
|
+
options.slice(
|
12
|
+
:delimiter, :unique_key
|
13
|
+
).reverse_merge(
|
14
|
+
delimiter: ',',
|
15
|
+
unique_key: [primary_key],
|
16
|
+
)
|
17
|
+
end
|
18
|
+
|
19
|
+
def database_connection
|
20
|
+
@destination.connection
|
21
|
+
end
|
22
|
+
|
23
|
+
def primary_key
|
24
|
+
@destination.primary_key
|
25
|
+
end
|
26
|
+
|
27
|
+
def column_names
|
28
|
+
@destination.column_names
|
29
|
+
end
|
30
|
+
|
31
|
+
def quoted_table_name
|
32
|
+
@destination.quoted_table_name
|
33
|
+
end
|
34
|
+
end
|
35
|
+
end
|
36
|
+
end
|
@@ -0,0 +1,56 @@
|
|
1
|
+
module PostgresUpsert
|
2
|
+
module WriteAdapters
|
3
|
+
class TableAdapter
|
4
|
+
def initialize(destination, options)
|
5
|
+
@destination = destination
|
6
|
+
@options = sanitize_options(options)
|
7
|
+
end
|
8
|
+
|
9
|
+
def sanitize_options(options)
|
10
|
+
options.slice(
|
11
|
+
:delimiter, :unique_key
|
12
|
+
).reverse_merge(
|
13
|
+
delimiter: ',',
|
14
|
+
unique_key: [primary_key],
|
15
|
+
)
|
16
|
+
end
|
17
|
+
|
18
|
+
def database_connection
|
19
|
+
ActiveRecord::Base.connection
|
20
|
+
end
|
21
|
+
|
22
|
+
def primary_key
|
23
|
+
@primary_key ||= begin
|
24
|
+
query = <<-SELECT_KEY
|
25
|
+
SELECT
|
26
|
+
pg_attribute.attname,
|
27
|
+
format_type(pg_attribute.atttypid, pg_attribute.atttypmod)
|
28
|
+
FROM pg_index, pg_class, pg_attribute
|
29
|
+
WHERE
|
30
|
+
pg_class.oid = '#{@destination}'::regclass AND
|
31
|
+
indrelid = pg_class.oid AND
|
32
|
+
pg_attribute.attrelid = pg_class.oid AND
|
33
|
+
pg_attribute.attnum = any(pg_index.indkey)
|
34
|
+
AND indisprimary
|
35
|
+
SELECT_KEY
|
36
|
+
|
37
|
+
pg_result = ActiveRecord::Base.connection.execute query
|
38
|
+
pg_result.each { |row| return row['attname'] }
|
39
|
+
end
|
40
|
+
end
|
41
|
+
|
42
|
+
def column_names
|
43
|
+
@column_names ||= begin
|
44
|
+
query = "SELECT * FROM information_schema.columns WHERE TABLE_NAME = '#{@destination}'"
|
45
|
+
pg_result = ActiveRecord::Base.connection.execute query
|
46
|
+
pg_result.map { |row| row['column_name'] }
|
47
|
+
end
|
48
|
+
end
|
49
|
+
|
50
|
+
def quoted_table_name
|
51
|
+
@quoted_table_name ||= database_connection.quote_table_name(@destination)
|
52
|
+
end
|
53
|
+
|
54
|
+
end
|
55
|
+
end
|
56
|
+
end
|