kiba-plus 0.1.2 → 0.1.3
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/.gitignore +3 -0
- data/.travis.yml +22 -1
- data/README.md +34 -9
- data/examples/customer_csv_to_mysql.etl +1 -1
- data/examples/customer_mysql_to_csv.etl +1 -1
- data/kiba-plus.gemspec +5 -0
- data/lib/kiba/plus/destination/csv.rb +21 -18
- data/lib/kiba/plus/destination/mysql.rb +8 -3
- data/lib/kiba/plus/destination/mysql_bulk.rb +70 -43
- data/lib/kiba/plus/destination/pg.rb +9 -2
- data/lib/kiba/plus/destination/pg_bulk.rb +72 -30
- data/lib/kiba/plus/destination/pg_bulk2.rb +69 -35
- data/lib/kiba/plus/destination/pg_bulk_utils.rb +44 -11
- data/lib/kiba/plus/helper.rb +9 -2
- data/lib/kiba/plus/job.rb +1 -1
- data/lib/kiba/plus/source/mysql.rb +10 -22
- data/lib/kiba/plus/source/pg.rb +40 -0
- data/lib/kiba/plus/version.rb +1 -1
- metadata +46 -3
checksums.yaml
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
---
|
2
2
|
SHA1:
|
3
|
-
metadata.gz:
|
4
|
-
data.tar.gz:
|
3
|
+
metadata.gz: b37f1fbad737a5141060feae03a0d99b98b20d42
|
4
|
+
data.tar.gz: f72a505eaf75bf1f24be63de334b63ecd955c881
|
5
5
|
SHA512:
|
6
|
-
metadata.gz:
|
7
|
-
data.tar.gz:
|
6
|
+
metadata.gz: 6f22ebbaa54bcf735a58235c9485342763aa8bcff593d292b6d055897db79a9963ffae44d71bdd9612f20cce76de77d2f6dbcdb04c0fc2b8369fd33734cf2f98
|
7
|
+
data.tar.gz: e7088d77e6e03bb90445f0424a72532fc264ef87503f17fb7e4468bd7d481de3b17fc9994c06b4aa042d66dcbde00c40cf6d68497748028e3c2102b7e61ba1dc
|
data/.gitignore
CHANGED
data/.travis.yml
CHANGED
@@ -1,4 +1,25 @@
|
|
1
1
|
language: ruby
|
2
|
+
|
2
3
|
rvm:
|
3
4
|
- 2.2.4
|
4
|
-
|
5
|
+
|
6
|
+
services:
|
7
|
+
- mysql
|
8
|
+
- postgresql
|
9
|
+
|
10
|
+
env:
|
11
|
+
- MYSQL2_SRC_CONNECT_URL=mysql2://travis@localhost/kiba_plus_src_test
|
12
|
+
- MYSQL2_DEST_CONNECT_URL=mysql2://travis@localhost/kiba_plus_dest_test
|
13
|
+
- PG_SRC_CONNECT_URL=postgresql://postgres@localhost/kiba_plus_src_test
|
14
|
+
- PG_DEST_CONNECT_URL=postgresql://postgres@localhost/kiba_plus_dest_test
|
15
|
+
|
16
|
+
before_install:
|
17
|
+
- gem install bundler -v 1.11.2
|
18
|
+
|
19
|
+
before_script:
|
20
|
+
- mysql -e 'create database kiba_plus_src_test;'
|
21
|
+
- mysql -e 'create database kiba_plus_dest_test;'
|
22
|
+
- psql -c 'create database kiba_plus_src_test;' -U postgres
|
23
|
+
- psql -c 'create database kiba_plus_dest_test;' -U postgres
|
24
|
+
|
25
|
+
script: bundle exec rake test
|
data/README.md
CHANGED
@@ -4,6 +4,8 @@ Kiba enhancement for Ruby ETL. It connects to various data sources including rel
|
|
4
4
|
# Usage
|
5
5
|
|
6
6
|
```ruby
|
7
|
+
# /tmp/customer_mysql_to_pg.etl
|
8
|
+
|
7
9
|
require 'kiba/plus'
|
8
10
|
|
9
11
|
SOURCE_URL = 'mysql://root@localhost/shopperplus'
|
@@ -27,22 +29,33 @@ post_process do
|
|
27
29
|
end
|
28
30
|
```
|
29
31
|
|
30
|
-
Execute:
|
32
|
+
Execute in shell:
|
31
33
|
|
32
34
|
```shell
|
33
|
-
bundle exec kiba customer_mysql_to_pg.etl
|
34
|
-
```
|
35
|
-
|
36
|
-
Output:
|
35
|
+
$ bundle exec kiba /tmp/customer_mysql_to_pg.etl
|
37
36
|
|
38
|
-
```
|
39
37
|
# Output:
|
40
38
|
# I, [2016-05-16T01:53:36.832565 #87909] INFO -- : TRUNCATE TABLE customers;
|
41
39
|
# I, [2016-05-16T01:53:36.841770 #87909] INFO -- : COPY customers (id, email, first_name, last_name) FROM STDIN WITH DELIMITER ',' NULL '\N' CSV
|
42
40
|
# Insert total: 428972
|
43
41
|
```
|
44
42
|
|
45
|
-
|
43
|
+
Execute in ruby script:
|
44
|
+
|
45
|
+
```ruby
|
46
|
+
require 'kiba'
|
47
|
+
|
48
|
+
job_definition = Kiba.parse(IO.read('/tmp/customer_mysql_to_pg.etl'), '/tmp/customer_mysql_to_pg.etl')
|
49
|
+
Kiba.run(job_definition)
|
50
|
+
```
|
51
|
+
|
52
|
+
# Examples
|
53
|
+
|
54
|
+
* [CSV to MySQL](https://github.com/hooopo/kiba-plus/blob/master/examples/customer_csv_to_mysql.etl)
|
55
|
+
* [CSV to PG](https://github.com/hooopo/kiba-plus/blob/master/examples/customer_csv_to_pg.etl)
|
56
|
+
* [MySQL to CSV](https://github.com/hooopo/kiba-plus/blob/master/examples/customer_mysql_to_csv.etl)
|
57
|
+
* [MySQL to PG](https://github.com/hooopo/kiba-plus/blob/master/examples/customer_mysql_to_pg.etl)
|
58
|
+
* [MySQL incremental to PG](https://github.com/hooopo/kiba-plus/blob/master/examples/incremental_insert.etl)
|
46
59
|
|
47
60
|
# Main Feature
|
48
61
|
|
@@ -73,6 +86,7 @@ More Examples(TODO).
|
|
73
86
|
Add this line to your application's Gemfile:
|
74
87
|
|
75
88
|
```ruby
|
89
|
+
gem 'kiba'
|
76
90
|
gem 'kiba-plus'
|
77
91
|
```
|
78
92
|
|
@@ -84,10 +98,21 @@ Or install it yourself as:
|
|
84
98
|
|
85
99
|
$ gem install kiba-plus
|
86
100
|
|
87
|
-
##
|
101
|
+
## Development
|
88
102
|
|
103
|
+
First of all, Please run the following code in shell.
|
89
104
|
|
90
|
-
|
105
|
+
```bash
|
106
|
+
|
107
|
+
$ mysql -e 'create database kiba_plus_src_test;'
|
108
|
+
|
109
|
+
$ mysql -e 'create database kiba_plus_dest_test;'
|
110
|
+
|
111
|
+
$ psql -c 'create database kiba_plus_src_test;' -U postgres
|
112
|
+
|
113
|
+
$ psql -c 'create database kiba_plus_dest_test;' -U postgres
|
114
|
+
|
115
|
+
```
|
91
116
|
|
92
117
|
After checking out the repo, run `bin/setup` to install dependencies. Then, run `rake test` to run the tests. You can also run `bin/console` for an interactive prompt that will allow you to experiment.
|
93
118
|
|
@@ -11,6 +11,6 @@ destination Kiba::Plus::Destination::MysqlBulk, { :connect_url => DEST_URL,
|
|
11
11
|
:incremental => false
|
12
12
|
}
|
13
13
|
post_process do
|
14
|
-
result = Mysql2::Client.new(
|
14
|
+
result = Mysql2::Client.new(mysql2_connect_hash(DEST_URL)).query("SELECT COUNT(*) AS num FROM customers")
|
15
15
|
puts "Insert total: #{result.first['num']}"
|
16
16
|
end
|
@@ -1,7 +1,7 @@
|
|
1
1
|
#!/usr/bin/env ruby
|
2
2
|
require_relative 'init'
|
3
3
|
|
4
|
-
|
4
|
+
SOURCE_URL = 'mysql://root@localhost/shopperplus'
|
5
5
|
|
6
6
|
source Kiba::Plus::Source::Mysql, :connect_url => SOURCE_URL,
|
7
7
|
:query => %Q{SELECT id, email, 'hooopo' AS first_name, 'Wang' AS last_name FROM customers}
|
data/kiba-plus.gemspec
CHANGED
@@ -26,10 +26,15 @@ Gem::Specification.new do |spec|
|
|
26
26
|
spec.bindir = "exe"
|
27
27
|
spec.executables = spec.files.grep(%r{^exe/}) { |f| File.basename(f) }
|
28
28
|
spec.require_paths = ["lib"]
|
29
|
+
|
29
30
|
spec.add_runtime_dependency "kiba", "~> 0.6"
|
30
31
|
spec.add_runtime_dependency "mysql2", "~> 0.4"
|
31
32
|
spec.add_runtime_dependency "pg", "~> 0.18"
|
33
|
+
|
32
34
|
spec.add_development_dependency "bundler", "~> 1.11"
|
33
35
|
spec.add_development_dependency "rake", "~> 10.0"
|
34
36
|
spec.add_development_dependency "minitest", "~> 5.0"
|
37
|
+
spec.add_development_dependency 'database_cleaner', '~> 1.5.3'
|
38
|
+
spec.add_development_dependency 'sequel', '~> 4.34'
|
39
|
+
spec.add_development_dependency 'pry'
|
35
40
|
end
|
@@ -2,17 +2,17 @@ require 'csv'
|
|
2
2
|
|
3
3
|
module Kiba::Plus::Destination
|
4
4
|
class Csv
|
5
|
-
attr_reader :options
|
5
|
+
attr_reader :options, :csv
|
6
6
|
|
7
7
|
def initialize(options = {})
|
8
8
|
@options = options
|
9
9
|
@options.assert_valid_keys(
|
10
10
|
:output_file,
|
11
|
+
:mode,
|
11
12
|
:row_sep,
|
12
13
|
:col_sep,
|
13
14
|
:force_quotes,
|
14
|
-
:quote_char
|
15
|
-
:mode
|
15
|
+
:quote_char
|
16
16
|
)
|
17
17
|
@csv = CSV.open(output_file, mode, {
|
18
18
|
:col_sep => col_sep,
|
@@ -22,36 +22,39 @@ module Kiba::Plus::Destination
|
|
22
22
|
})
|
23
23
|
end
|
24
24
|
|
25
|
-
def
|
26
|
-
|
25
|
+
def write(row)
|
26
|
+
@csv << row.values
|
27
27
|
end
|
28
28
|
|
29
|
-
def
|
30
|
-
|
29
|
+
def close
|
30
|
+
@csv.close
|
31
31
|
end
|
32
32
|
|
33
|
-
|
34
|
-
options.fetch(:col_sep, ",")
|
35
|
-
end
|
33
|
+
private
|
36
34
|
|
37
|
-
def
|
38
|
-
options.fetch(:
|
35
|
+
def output_file
|
36
|
+
options.fetch(:output_file)
|
39
37
|
end
|
40
38
|
|
41
|
-
def
|
42
|
-
options.fetch(:
|
39
|
+
def mode
|
40
|
+
options.fetch(:mode, "w")
|
43
41
|
end
|
44
42
|
|
45
43
|
def row_sep
|
46
44
|
options.fetch(:row_sep, "\n")
|
47
45
|
end
|
48
46
|
|
49
|
-
def
|
50
|
-
|
47
|
+
def col_sep
|
48
|
+
options.fetch(:col_sep, ",")
|
51
49
|
end
|
52
50
|
|
53
|
-
def
|
54
|
-
|
51
|
+
def force_quotes
|
52
|
+
options.fetch(:force_quotes, false)
|
53
|
+
end
|
54
|
+
|
55
|
+
def quote_char
|
56
|
+
options.fetch(:quote_char, '"')
|
55
57
|
end
|
58
|
+
|
56
59
|
end
|
57
60
|
end
|
@@ -3,7 +3,7 @@ require 'mysql2'
|
|
3
3
|
module Kiba::Plus::Destination
|
4
4
|
class Mysql
|
5
5
|
include Kiba::Plus::Helper
|
6
|
-
attr_reader :options
|
6
|
+
attr_reader :options, :client
|
7
7
|
|
8
8
|
def initialize(options = {})
|
9
9
|
@options = options
|
@@ -12,8 +12,8 @@ module Kiba::Plus::Destination
|
|
12
12
|
:prepare_sql,
|
13
13
|
:columns
|
14
14
|
)
|
15
|
-
@client = Mysql2::Client.new(
|
16
|
-
|
15
|
+
@client = Mysql2::Client.new(mysql2_connect_hash(connect_url))
|
16
|
+
init
|
17
17
|
end
|
18
18
|
|
19
19
|
def write(row)
|
@@ -21,6 +21,7 @@ module Kiba::Plus::Destination
|
|
21
21
|
rescue => e
|
22
22
|
Kiba::Plus.logger.error "ERROR for #{row}"
|
23
23
|
Kiba::Plus.logger.error e.message
|
24
|
+
raise e
|
24
25
|
end
|
25
26
|
|
26
27
|
def close
|
@@ -30,6 +31,10 @@ module Kiba::Plus::Destination
|
|
30
31
|
|
31
32
|
private
|
32
33
|
|
34
|
+
def init
|
35
|
+
@pre_stmt = @client.prepare(prepare_sql)
|
36
|
+
end
|
37
|
+
|
33
38
|
def connect_url
|
34
39
|
options.fetch(:connect_url)
|
35
40
|
end
|
@@ -3,45 +3,53 @@ require 'mysql2'
|
|
3
3
|
module Kiba::Plus::Destination
|
4
4
|
class MysqlBulk
|
5
5
|
include Kiba::Plus::Helper
|
6
|
-
attr_reader :options
|
6
|
+
attr_reader :options, :client
|
7
7
|
|
8
8
|
def initialize(options = {})
|
9
9
|
@options = options
|
10
|
-
@options.assert_valid_keys(
|
11
|
-
|
12
|
-
|
13
|
-
|
14
|
-
|
15
|
-
|
16
|
-
|
17
|
-
|
18
|
-
|
19
|
-
|
20
|
-
|
21
|
-
|
10
|
+
@options.assert_valid_keys(
|
11
|
+
:connect_url,
|
12
|
+
:table_name,
|
13
|
+
:columns,
|
14
|
+
:truncate,
|
15
|
+
:incremental,
|
16
|
+
:input_file,
|
17
|
+
:ignore_input_file_header,
|
18
|
+
:delimited_by,
|
19
|
+
:enclosed_by,
|
20
|
+
:ignore_lines
|
21
|
+
)
|
22
|
+
|
23
|
+
@client = Mysql2::Client.new(mysql2_connect_hash(connect_url).merge(local_infile: true))
|
22
24
|
end
|
23
25
|
|
24
|
-
def
|
25
|
-
|
26
|
+
def write(row)
|
27
|
+
# blank!
|
26
28
|
end
|
27
29
|
|
28
|
-
def
|
29
|
-
|
30
|
-
|
30
|
+
def close
|
31
|
+
if truncate
|
32
|
+
sql = truncate_sql
|
33
|
+
Kiba::Plus.logger.info sql
|
34
|
+
client.query(sql)
|
35
|
+
end
|
31
36
|
|
32
|
-
|
33
|
-
|
34
|
-
|
37
|
+
sql = bulk_sql
|
38
|
+
Kiba::Plus.logger.info sql
|
39
|
+
client.query(sql)
|
35
40
|
|
36
|
-
|
37
|
-
|
41
|
+
client.close
|
42
|
+
@client = nil
|
38
43
|
end
|
39
44
|
|
40
|
-
|
41
|
-
|
45
|
+
private
|
46
|
+
|
47
|
+
def connect_url
|
48
|
+
options.fetch(:connect_url)
|
42
49
|
end
|
43
50
|
|
44
|
-
def
|
51
|
+
def table_name
|
52
|
+
options.fetch(:table_name)
|
45
53
|
end
|
46
54
|
|
47
55
|
def columns
|
@@ -60,27 +68,46 @@ module Kiba::Plus::Destination
|
|
60
68
|
options.fetch(:input_file)
|
61
69
|
end
|
62
70
|
|
63
|
-
def
|
64
|
-
|
65
|
-
|
66
|
-
|
67
|
-
|
68
|
-
|
71
|
+
def ignore_input_file_header
|
72
|
+
!!options.fetch(:ignore_input_file_header, false)
|
73
|
+
end
|
74
|
+
|
75
|
+
def delimited_by
|
76
|
+
options.fetch(:delimited_by, ",")
|
77
|
+
end
|
78
|
+
|
79
|
+
def enclosed_by
|
80
|
+
options.fetch(:enclosed_by, '"')
|
81
|
+
end
|
82
|
+
|
83
|
+
def ignore_lines
|
84
|
+
options.fetch(:ignore_lines, 0).to_i
|
85
|
+
end
|
86
|
+
|
87
|
+
def real_ignore_lines
|
88
|
+
lines = ignore_lines
|
89
|
+
lines += 1 if ignore_input_file_header
|
90
|
+
lines
|
91
|
+
end
|
69
92
|
|
70
|
-
|
71
|
-
|
72
|
-
|
73
|
-
|
93
|
+
def truncate_sql
|
94
|
+
sql = "TRUNCATE TABLE #{table_name}"
|
95
|
+
format_sql sql
|
96
|
+
end
|
97
|
+
|
98
|
+
def bulk_sql
|
99
|
+
sql = <<-SQL
|
100
|
+
LOAD DATA LOCAL INFILE '#{input_file}'
|
101
|
+
REPLACE
|
102
|
+
INTO TABLE #{table_name}
|
103
|
+
FIELDS
|
74
104
|
TERMINATED BY '#{delimited_by}'
|
75
105
|
ENCLOSED BY '#{enclosed_by}'
|
76
|
-
|
77
|
-
|
106
|
+
IGNORE #{real_ignore_lines} LINES
|
107
|
+
(#{columns.join(',')})
|
78
108
|
SQL
|
79
|
-
|
80
|
-
@client.query(bulk_sql)
|
81
|
-
|
82
|
-
@client.close
|
83
|
-
@client = nil
|
109
|
+
format_sql sql
|
84
110
|
end
|
111
|
+
|
85
112
|
end
|
86
113
|
end
|
@@ -2,18 +2,20 @@ require 'pg'
|
|
2
2
|
|
3
3
|
module Kiba::Plus::Destination
|
4
4
|
class Pg
|
5
|
-
attr_reader :options
|
5
|
+
attr_reader :options, :conn
|
6
6
|
|
7
7
|
def initialize(options = {})
|
8
8
|
@options = options
|
9
9
|
@options.assert_valid_keys(
|
10
10
|
:connect_url,
|
11
|
+
:schema,
|
11
12
|
:prepare_name,
|
12
13
|
:prepare_sql,
|
13
14
|
:columns
|
14
15
|
)
|
15
16
|
@conn = PG.connect(connect_url)
|
16
|
-
@conn.
|
17
|
+
@conn.exec "SET search_path TO %s" % [ options.fetch(:schema) ] unless options.fetch(:schema).empty?
|
18
|
+
init
|
17
19
|
end
|
18
20
|
|
19
21
|
def write(row)
|
@@ -23,6 +25,7 @@ module Kiba::Plus::Destination
|
|
23
25
|
Kiba::Plus.logger.error "ERROR for #{row}"
|
24
26
|
Kiba::Plus.logger.error ex.message
|
25
27
|
# Maybe, write to db table or file
|
28
|
+
raise ex
|
26
29
|
end
|
27
30
|
|
28
31
|
def close
|
@@ -32,6 +35,10 @@ module Kiba::Plus::Destination
|
|
32
35
|
|
33
36
|
private
|
34
37
|
|
38
|
+
def init
|
39
|
+
@conn.prepare(prepare_name, prepare_sql)
|
40
|
+
end
|
41
|
+
|
35
42
|
def connect_url
|
36
43
|
options.fetch(:connect_url)
|
37
44
|
end
|
@@ -2,20 +2,48 @@ require_relative 'pg_bulk_utils'
|
|
2
2
|
module Kiba::Plus::Destination
|
3
3
|
class PgBulk
|
4
4
|
include PgBulkUtils
|
5
|
-
|
5
|
+
include Kiba::Plus::Helper
|
6
|
+
attr_reader :options, :conn
|
6
7
|
|
7
8
|
def initialize(options = {})
|
8
9
|
@options = options
|
9
10
|
@options.assert_valid_keys(
|
10
|
-
|
11
|
-
|
12
|
-
|
13
|
-
|
14
|
-
|
15
|
-
|
16
|
-
|
17
|
-
|
11
|
+
:connect_url,
|
12
|
+
:table_name,
|
13
|
+
:columns,
|
14
|
+
:truncate,
|
15
|
+
:incremental,
|
16
|
+
:unique_by,
|
17
|
+
:input_file,
|
18
|
+
:ignore_input_file_header
|
19
|
+
)
|
18
20
|
@conn = PG.connect(connect_url)
|
21
|
+
|
22
|
+
init
|
23
|
+
end
|
24
|
+
|
25
|
+
def write(row)
|
26
|
+
# blank!
|
27
|
+
end
|
28
|
+
|
29
|
+
def close
|
30
|
+
if incremental
|
31
|
+
truncate_staging_table
|
32
|
+
create_staging_table
|
33
|
+
copy_to_staging_table
|
34
|
+
delete_before_insert
|
35
|
+
merge_to_target_table
|
36
|
+
truncate_staging_table
|
37
|
+
else
|
38
|
+
copy_to_target_table
|
39
|
+
end
|
40
|
+
@conn.close
|
41
|
+
@conn = nil
|
42
|
+
end
|
43
|
+
|
44
|
+
private
|
45
|
+
|
46
|
+
def init
|
19
47
|
if truncate
|
20
48
|
truncate_staging_table
|
21
49
|
truncate_target_table
|
@@ -30,10 +58,6 @@ module Kiba::Plus::Destination
|
|
30
58
|
options.fetch(:table_name)
|
31
59
|
end
|
32
60
|
|
33
|
-
def input_file
|
34
|
-
options.fetch(:input_file)
|
35
|
-
end
|
36
|
-
|
37
61
|
def columns
|
38
62
|
options.fetch(:columns)
|
39
63
|
end
|
@@ -50,35 +74,53 @@ module Kiba::Plus::Destination
|
|
50
74
|
options.fetch(:unique_by, :id)
|
51
75
|
end
|
52
76
|
|
77
|
+
def input_file
|
78
|
+
options.fetch(:input_file)
|
79
|
+
end
|
80
|
+
|
81
|
+
def ignore_input_file_header
|
82
|
+
!!options.fetch(:ignore_input_file_header, false)
|
83
|
+
end
|
84
|
+
|
53
85
|
def copy_to_target_table
|
54
|
-
sql =
|
86
|
+
sql = copy_to_target_table_sql
|
55
87
|
Kiba::Plus.logger.info sql
|
56
88
|
@conn.exec(sql)
|
57
89
|
end
|
58
90
|
|
59
91
|
def copy_to_staging_table
|
60
|
-
sql =
|
92
|
+
sql = copy_to_staging_table_sql
|
61
93
|
Kiba::Plus.logger.info sql
|
62
94
|
@conn.exec(sql)
|
63
95
|
end
|
64
96
|
|
65
|
-
def
|
66
|
-
|
97
|
+
def copy_to_target_table_sql
|
98
|
+
sql = <<-SQL
|
99
|
+
COPY #{table_name} (#{columns.join(', ')})
|
100
|
+
FROM '#{File.expand_path(input_file)}'
|
101
|
+
WITH
|
102
|
+
#{ignore_input_file_header ? 'HEADER' : ''}
|
103
|
+
DELIMITER ','
|
104
|
+
NULL '\\N'
|
105
|
+
CSV
|
106
|
+
SQL
|
107
|
+
|
108
|
+
format_sql sql
|
67
109
|
end
|
68
110
|
|
69
|
-
def
|
70
|
-
|
71
|
-
|
72
|
-
|
73
|
-
|
74
|
-
|
75
|
-
|
76
|
-
|
77
|
-
|
78
|
-
|
79
|
-
|
80
|
-
|
81
|
-
@conn = nil
|
111
|
+
def copy_to_staging_table_sql
|
112
|
+
sql = <<-SQL
|
113
|
+
COPY #{staging_table_name} (#{columns.join(', ')})
|
114
|
+
FROM '#{File.expand_path(input_file)}'
|
115
|
+
WITH
|
116
|
+
#{ignore_input_file_header ? 'HEADER' : ''}
|
117
|
+
DELIMITER ','
|
118
|
+
NULL '\\N'
|
119
|
+
CSV
|
120
|
+
SQL
|
121
|
+
|
122
|
+
format_sql sql
|
82
123
|
end
|
124
|
+
|
83
125
|
end
|
84
126
|
end
|
@@ -4,19 +4,54 @@ require_relative 'pg_bulk_utils'
|
|
4
4
|
module Kiba::Plus::Destination
|
5
5
|
class PgBulk2
|
6
6
|
include PgBulkUtils
|
7
|
-
|
7
|
+
include Kiba::Plus::Helper
|
8
|
+
attr_reader :options, :conn
|
8
9
|
|
9
10
|
def initialize(options = {})
|
10
11
|
@options = options
|
11
|
-
@options.assert_valid_keys(
|
12
|
-
|
13
|
-
|
14
|
-
|
15
|
-
|
16
|
-
|
17
|
-
|
12
|
+
@options.assert_valid_keys(
|
13
|
+
:connect_url,
|
14
|
+
:table_name,
|
15
|
+
:columns,
|
16
|
+
:truncate,
|
17
|
+
:incremental,
|
18
|
+
:unique_by
|
19
|
+
)
|
18
20
|
|
19
21
|
@conn = PG.connect(connect_url)
|
22
|
+
|
23
|
+
init
|
24
|
+
end
|
25
|
+
|
26
|
+
def write(row)
|
27
|
+
begin
|
28
|
+
@conn.put_copy_data CSV.generate_line(row.values_at(*columns))
|
29
|
+
rescue Exception => err
|
30
|
+
errmsg = "%s while copy data: %s" % [ err.class.name, err.message ]
|
31
|
+
@conn.put_copy_end( errmsg )
|
32
|
+
Kiba::Plus.logger.error @conn.get_result
|
33
|
+
raise
|
34
|
+
end
|
35
|
+
end
|
36
|
+
|
37
|
+
def close
|
38
|
+
@conn.put_copy_end
|
39
|
+
@conn.get_last_result
|
40
|
+
if incremental
|
41
|
+
delete_before_insert
|
42
|
+
merge_to_target_table
|
43
|
+
truncate_staging_table
|
44
|
+
end
|
45
|
+
rescue
|
46
|
+
raise
|
47
|
+
ensure
|
48
|
+
@conn.close
|
49
|
+
@conn = nil
|
50
|
+
end
|
51
|
+
|
52
|
+
private
|
53
|
+
|
54
|
+
def init
|
20
55
|
if truncate
|
21
56
|
truncate_staging_table
|
22
57
|
truncate_target_table
|
@@ -24,12 +59,12 @@ module Kiba::Plus::Destination
|
|
24
59
|
if incremental
|
25
60
|
truncate_staging_table
|
26
61
|
create_staging_table
|
27
|
-
sql =
|
62
|
+
sql = bulk_sql_with_incremental
|
28
63
|
else
|
29
|
-
sql =
|
64
|
+
sql = bulk_sql_with_non_incremental
|
30
65
|
end
|
31
66
|
Kiba::Plus.logger.info sql
|
32
|
-
@
|
67
|
+
@conn.exec(sql)
|
33
68
|
end
|
34
69
|
|
35
70
|
def connect_url
|
@@ -40,17 +75,6 @@ module Kiba::Plus::Destination
|
|
40
75
|
options.fetch(:table_name)
|
41
76
|
end
|
42
77
|
|
43
|
-
def write(row)
|
44
|
-
begin
|
45
|
-
@conn.put_copy_data CSV.generate_line(row.values_at(*columns))
|
46
|
-
rescue Exception => err
|
47
|
-
errmsg = "%s while copy data: %s" % [ err.class.name, err.message ]
|
48
|
-
@conn.put_copy_end( errmsg )
|
49
|
-
Kiba::Plus.logger.error @conn.get_result
|
50
|
-
raise
|
51
|
-
end
|
52
|
-
end
|
53
|
-
|
54
78
|
def columns
|
55
79
|
options.fetch(:columns)
|
56
80
|
end
|
@@ -67,19 +91,29 @@ module Kiba::Plus::Destination
|
|
67
91
|
options.fetch(:unique_by, :id)
|
68
92
|
end
|
69
93
|
|
70
|
-
def
|
71
|
-
|
72
|
-
|
73
|
-
|
74
|
-
|
75
|
-
|
76
|
-
|
77
|
-
|
78
|
-
|
79
|
-
|
80
|
-
ensure
|
81
|
-
@conn.close
|
82
|
-
@conn = nil
|
94
|
+
def bulk_sql_with_incremental
|
95
|
+
sql = <<-SQL
|
96
|
+
COPY #{staging_table_name} (#{columns.join(', ')})
|
97
|
+
FROM STDIN
|
98
|
+
WITH
|
99
|
+
DELIMITER ','
|
100
|
+
NULL '\\N'
|
101
|
+
CSV
|
102
|
+
SQL
|
103
|
+
format_sql sql
|
83
104
|
end
|
105
|
+
|
106
|
+
def bulk_sql_with_non_incremental
|
107
|
+
sql = <<-SQL
|
108
|
+
COPY #{table_name} (#{columns.join(', ')})
|
109
|
+
FROM STDIN
|
110
|
+
WITH
|
111
|
+
DELIMITER ','
|
112
|
+
NULL '\\N'
|
113
|
+
CSV
|
114
|
+
SQL
|
115
|
+
format_sql sql
|
116
|
+
end
|
117
|
+
|
84
118
|
end
|
85
119
|
end
|
@@ -1,39 +1,72 @@
|
|
1
1
|
module Kiba::Plus::Destination
|
2
2
|
module PgBulkUtils
|
3
|
+
|
4
|
+
private
|
5
|
+
|
3
6
|
def staging_table_name
|
4
7
|
table_name + "_staging"
|
5
8
|
end
|
6
9
|
|
7
10
|
def create_staging_table
|
8
|
-
sql =
|
11
|
+
sql = create_staging_table_sql
|
9
12
|
Kiba::Plus.logger.info sql
|
10
13
|
@conn.exec(sql)
|
11
14
|
end
|
12
15
|
|
16
|
+
def create_staging_table_sql
|
17
|
+
sql = <<-SQL
|
18
|
+
CREATE TABLE IF NOT EXISTS #{staging_table_name} (
|
19
|
+
LIKE #{table_name} INCLUDING DEFAULTS INCLUDING CONSTRAINTS INCLUDING INDEXES
|
20
|
+
)
|
21
|
+
SQL
|
22
|
+
format_sql sql
|
23
|
+
end
|
24
|
+
|
13
25
|
def truncate_staging_table
|
14
|
-
|
15
|
-
Kiba::Plus.logger.info
|
16
|
-
@conn.exec(
|
26
|
+
sql = truncate_staging_table_sql
|
27
|
+
Kiba::Plus.logger.info sql
|
28
|
+
@conn.exec(sql) rescue nil
|
29
|
+
end
|
30
|
+
|
31
|
+
def truncate_staging_table_sql
|
32
|
+
sql = "TRUNCATE TABLE #{staging_table_name}"
|
33
|
+
format_sql sql
|
17
34
|
end
|
18
35
|
|
19
36
|
def truncate_target_table
|
20
|
-
|
21
|
-
Kiba::Plus.logger.info
|
22
|
-
@conn.exec(
|
37
|
+
sql = truncate_target_table_sql
|
38
|
+
Kiba::Plus.logger.info sql
|
39
|
+
@conn.exec(sql)
|
40
|
+
end
|
41
|
+
|
42
|
+
def truncate_target_table_sql
|
43
|
+
sql = "TRUNCATE TABLE #{table_name}"
|
44
|
+
format_sql sql
|
23
45
|
end
|
24
46
|
|
25
|
-
# TODO add where condition to speed up deleting.
|
26
47
|
def delete_before_insert
|
27
|
-
|
28
|
-
sql = "DELETE FROM #{table_name} USING #{staging_table_name} WHERE #{where}"
|
48
|
+
sql = delete_before_insert_sql
|
29
49
|
Kiba::Plus.logger.info sql
|
30
50
|
@conn.exec(sql)
|
31
51
|
end
|
32
52
|
|
53
|
+
# TODO add where condition to speed up deleting.
|
54
|
+
def delete_before_insert_sql
|
55
|
+
where = Array(unique_by).map{|x| ["#{staging_table_name}.#{x}", "#{table_name}.#{x}"].join(" = ") }.join(" AND ")
|
56
|
+
sql = "DELETE FROM #{table_name} USING #{staging_table_name} WHERE #{where}"
|
57
|
+
format_sql sql
|
58
|
+
end
|
59
|
+
|
33
60
|
def merge_to_target_table
|
34
|
-
sql =
|
61
|
+
sql = merge_to_target_table_sql
|
35
62
|
Kiba::Plus.logger.info sql
|
36
63
|
@conn.exec(sql)
|
37
64
|
end
|
65
|
+
|
66
|
+
def merge_to_target_table_sql
|
67
|
+
sql = "INSERT INTO #{table_name} (SELECT * FROM #{staging_table_name})"
|
68
|
+
format_sql sql
|
69
|
+
end
|
70
|
+
|
38
71
|
end
|
39
72
|
end
|
data/lib/kiba/plus/helper.rb
CHANGED
@@ -2,13 +2,15 @@ require 'uri'
|
|
2
2
|
module Kiba
|
3
3
|
module Plus
|
4
4
|
module Helper
|
5
|
-
def
|
5
|
+
def mysql2_connect_hash(url)
|
6
|
+
return url if url.is_a?(Hash)
|
7
|
+
|
6
8
|
u = URI.parse(url)
|
7
9
|
{
|
8
10
|
host: u.host,
|
11
|
+
port: u.port,
|
9
12
|
username: u.user,
|
10
13
|
password: u.password,
|
11
|
-
port: u.port,
|
12
14
|
database: u.path[1..-1]
|
13
15
|
}
|
14
16
|
end
|
@@ -17,6 +19,11 @@ module Kiba
|
|
17
19
|
u = URI.parse(url)
|
18
20
|
u.scheme
|
19
21
|
end
|
22
|
+
|
23
|
+
def format_sql(sql)
|
24
|
+
sql.to_s.gsub(/[\n][\s]*[\n]/, "\n")
|
25
|
+
end
|
26
|
+
|
20
27
|
end
|
21
28
|
end
|
22
29
|
end
|
data/lib/kiba/plus/job.rb
CHANGED
@@ -15,7 +15,7 @@ module Kiba
|
|
15
15
|
@options.assert_valid_keys(:connect_url, :job_id, :job_name, :start_at, :completed_at)
|
16
16
|
url = URI.parse(connect_url)
|
17
17
|
if url.scheme =~ /mysql/i
|
18
|
-
@client = Mysql2::Client.new(
|
18
|
+
@client = Mysql2::Client.new(mysql2_connect_hash(connect_url))
|
19
19
|
elsif url.scheme =~ /postgres/i
|
20
20
|
@client = PG.connect(connect_url)
|
21
21
|
else
|
@@ -10,41 +10,29 @@ module Kiba
|
|
10
10
|
def initialize(options = {})
|
11
11
|
@options = options
|
12
12
|
@options.assert_valid_keys(
|
13
|
-
:
|
14
|
-
:
|
15
|
-
|
16
|
-
|
17
|
-
:connect_url
|
18
|
-
)
|
19
|
-
@client = Mysql2::Client.new(connect_hash(connect_url))
|
13
|
+
:connect_url,
|
14
|
+
:query
|
15
|
+
)
|
16
|
+
@client = Mysql2::Client.new(mysql2_connect_hash(connect_url))
|
20
17
|
end
|
21
18
|
|
22
19
|
def each
|
20
|
+
Kiba::Plus.logger.info query
|
23
21
|
results = client.query(query, as: :hash, symbolize_keys: true, stream: true)
|
24
22
|
results.each do |row|
|
25
23
|
yield(row)
|
26
24
|
end
|
27
25
|
end
|
28
26
|
|
29
|
-
|
30
|
-
options.fetch(:query)
|
31
|
-
end
|
32
|
-
|
33
|
-
def output
|
34
|
-
options.fetch(:output)
|
35
|
-
end
|
36
|
-
|
37
|
-
def last_pull_at
|
38
|
-
options[:last_pull_at]
|
39
|
-
end
|
40
|
-
|
41
|
-
def incremental
|
42
|
-
options.fetch(:incremental, true)
|
43
|
-
end
|
27
|
+
private
|
44
28
|
|
45
29
|
def connect_url
|
46
30
|
options.fetch(:connect_url)
|
47
31
|
end
|
32
|
+
|
33
|
+
def query
|
34
|
+
options.fetch(:query)
|
35
|
+
end
|
48
36
|
end
|
49
37
|
end
|
50
38
|
end
|
@@ -0,0 +1,40 @@
|
|
1
|
+
require 'pg'
|
2
|
+
require 'uri'
|
3
|
+
|
4
|
+
module Kiba
|
5
|
+
module Plus::Source
|
6
|
+
class Pg
|
7
|
+
include Kiba::Plus::Helper
|
8
|
+
attr_reader :options, :client
|
9
|
+
|
10
|
+
def initialize(options = {})
|
11
|
+
@options = options
|
12
|
+
@options.assert_valid_keys(
|
13
|
+
:connect_url,
|
14
|
+
:schema,
|
15
|
+
:query
|
16
|
+
)
|
17
|
+
@client = PG.connect(connect_url)
|
18
|
+
@client.exec "SET search_path TO %s" % [ options.fetch(:schema) ] unless options.fetch(:schema).empty?
|
19
|
+
end
|
20
|
+
|
21
|
+
def each
|
22
|
+
Kiba::Plus.logger.info query
|
23
|
+
results = client.query(query)
|
24
|
+
results.each do |row|
|
25
|
+
yield(row)
|
26
|
+
end
|
27
|
+
end
|
28
|
+
|
29
|
+
private
|
30
|
+
|
31
|
+
def connect_url
|
32
|
+
options.fetch(:connect_url)
|
33
|
+
end
|
34
|
+
|
35
|
+
def query
|
36
|
+
options.fetch(:query)
|
37
|
+
end
|
38
|
+
end
|
39
|
+
end
|
40
|
+
end
|
data/lib/kiba/plus/version.rb
CHANGED
metadata
CHANGED
@@ -1,14 +1,14 @@
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
2
2
|
name: kiba-plus
|
3
3
|
version: !ruby/object:Gem::Version
|
4
|
-
version: 0.1.
|
4
|
+
version: 0.1.3
|
5
5
|
platform: ruby
|
6
6
|
authors:
|
7
7
|
- Hooopo
|
8
8
|
autorequire:
|
9
9
|
bindir: exe
|
10
10
|
cert_chain: []
|
11
|
-
date:
|
11
|
+
date: 2017-08-10 00:00:00.000000000 Z
|
12
12
|
dependencies:
|
13
13
|
- !ruby/object:Gem::Dependency
|
14
14
|
name: kiba
|
@@ -94,6 +94,48 @@ dependencies:
|
|
94
94
|
- - "~>"
|
95
95
|
- !ruby/object:Gem::Version
|
96
96
|
version: '5.0'
|
97
|
+
- !ruby/object:Gem::Dependency
|
98
|
+
name: database_cleaner
|
99
|
+
requirement: !ruby/object:Gem::Requirement
|
100
|
+
requirements:
|
101
|
+
- - "~>"
|
102
|
+
- !ruby/object:Gem::Version
|
103
|
+
version: 1.5.3
|
104
|
+
type: :development
|
105
|
+
prerelease: false
|
106
|
+
version_requirements: !ruby/object:Gem::Requirement
|
107
|
+
requirements:
|
108
|
+
- - "~>"
|
109
|
+
- !ruby/object:Gem::Version
|
110
|
+
version: 1.5.3
|
111
|
+
- !ruby/object:Gem::Dependency
|
112
|
+
name: sequel
|
113
|
+
requirement: !ruby/object:Gem::Requirement
|
114
|
+
requirements:
|
115
|
+
- - "~>"
|
116
|
+
- !ruby/object:Gem::Version
|
117
|
+
version: '4.34'
|
118
|
+
type: :development
|
119
|
+
prerelease: false
|
120
|
+
version_requirements: !ruby/object:Gem::Requirement
|
121
|
+
requirements:
|
122
|
+
- - "~>"
|
123
|
+
- !ruby/object:Gem::Version
|
124
|
+
version: '4.34'
|
125
|
+
- !ruby/object:Gem::Dependency
|
126
|
+
name: pry
|
127
|
+
requirement: !ruby/object:Gem::Requirement
|
128
|
+
requirements:
|
129
|
+
- - ">="
|
130
|
+
- !ruby/object:Gem::Version
|
131
|
+
version: '0'
|
132
|
+
type: :development
|
133
|
+
prerelease: false
|
134
|
+
version_requirements: !ruby/object:Gem::Requirement
|
135
|
+
requirements:
|
136
|
+
- - ">="
|
137
|
+
- !ruby/object:Gem::Version
|
138
|
+
version: '0'
|
97
139
|
description: It connects to various data sources including relational, non-relational,
|
98
140
|
and flat file, cloud services and HTTP resources. It has flexible load strategies
|
99
141
|
including insert, bulk load and upsert.
|
@@ -135,6 +177,7 @@ files:
|
|
135
177
|
- lib/kiba/plus/job.rb
|
136
178
|
- lib/kiba/plus/logger.rb
|
137
179
|
- lib/kiba/plus/source/mysql.rb
|
180
|
+
- lib/kiba/plus/source/pg.rb
|
138
181
|
- lib/kiba/plus/version.rb
|
139
182
|
homepage: https://github.com/hooopo/kiba-plus
|
140
183
|
licenses:
|
@@ -157,7 +200,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
|
|
157
200
|
version: '0'
|
158
201
|
requirements: []
|
159
202
|
rubyforge_project:
|
160
|
-
rubygems_version: 2.
|
203
|
+
rubygems_version: 2.6.11
|
161
204
|
signing_key:
|
162
205
|
specification_version: 4
|
163
206
|
summary: Kiba enhancement for Ruby ETL
|