parqueteur 1.1.0 → 1.3.1
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/Gemfile.lock +8 -8
- data/README.md +137 -8
- data/examples/cars.rb +40 -0
- data/examples/convert-and-compression.rb +56 -0
- data/examples/convert-methods.rb +54 -0
- data/examples/convert-to-io.rb +52 -0
- data/examples/convert-with-chunks.rb +54 -0
- data/examples/convert-without-compression.rb +52 -0
- data/examples/hello-world.rb +57 -0
- data/examples/readme-example.rb +44 -0
- data/lib/parqueteur/converter.rb +108 -75
- data/lib/parqueteur/input.rb +12 -27
- data/lib/parqueteur/type_resolver.rb +10 -0
- data/lib/parqueteur/types/date32_type.rb +15 -0
- data/lib/parqueteur/types/date64_type.rb +15 -0
- data/lib/parqueteur/types/decimal128_type.rb +29 -0
- data/lib/parqueteur/types/decimal256_type.rb +29 -0
- data/lib/parqueteur/types/int32_type.rb +0 -2
- data/lib/parqueteur/types/int64_type.rb +0 -2
- data/lib/parqueteur/types/time32_type.rb +19 -0
- data/lib/parqueteur/types/time64_type.rb +19 -0
- data/lib/parqueteur/types/timestamp_type.rb +3 -1
- data/lib/parqueteur/version.rb +1 -1
- data/lib/parqueteur.rb +9 -4
- data/parqueteur.gemspec +2 -2
- data/scripts/apache-arrow-ubuntu-install.sh +18 -0
- metadata +19 -7
- data/example.rb +0 -39
- data/lib/parqueteur/chunked_converter.rb +0 -28
- data/test.json +0 -1
checksums.yaml
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
---
|
2
2
|
SHA256:
|
3
|
-
metadata.gz:
|
4
|
-
data.tar.gz:
|
3
|
+
metadata.gz: 86f767a68e38cdd93da015e4ddfcda06b2eefe553ecb6d7a423b4cc0f2752183
|
4
|
+
data.tar.gz: cd741b1d023e44fcc14c08192b9791b675339736527f47dbe2393e85bdff9d07
|
5
5
|
SHA512:
|
6
|
-
metadata.gz:
|
7
|
-
data.tar.gz:
|
6
|
+
metadata.gz: 58882a4d2d1ea5a0cb53f5643f96ac504352b87e6c1436d10841a616d2867bf127c70a31374f6ff2f455fb2894d29fe8cc439ca8a0efca5fe519c00ee312c8c5
|
7
|
+
data.tar.gz: ac185e758d8c0fa19d11ac05e96f77f93cb8260e055d203d1d0d71dd04e0ff6db70d773824a8a6fe5c63e5f235ab179f3417ac689b590403b69192cbad0bde98
|
data/Gemfile.lock
CHANGED
@@ -1,21 +1,21 @@
|
|
1
1
|
PATH
|
2
2
|
remote: .
|
3
3
|
specs:
|
4
|
-
parqueteur (1.1
|
4
|
+
parqueteur (1.3.1)
|
5
5
|
red-parquet (~> 5.0)
|
6
6
|
|
7
7
|
GEM
|
8
8
|
remote: https://rubygems.org/
|
9
9
|
specs:
|
10
|
-
bigdecimal (3.0.
|
11
|
-
extpp (0.0
|
12
|
-
gio2 (3.4.
|
13
|
-
gobject-introspection (= 3.4.
|
14
|
-
glib2 (3.4.
|
10
|
+
bigdecimal (3.0.2)
|
11
|
+
extpp (0.1.0)
|
12
|
+
gio2 (3.4.9)
|
13
|
+
gobject-introspection (= 3.4.9)
|
14
|
+
glib2 (3.4.9)
|
15
15
|
native-package-installer (>= 1.0.3)
|
16
16
|
pkg-config (>= 1.3.5)
|
17
|
-
gobject-introspection (3.4.
|
18
|
-
glib2 (= 3.4.
|
17
|
+
gobject-introspection (3.4.9)
|
18
|
+
glib2 (= 3.4.9)
|
19
19
|
native-package-installer (1.1.1)
|
20
20
|
pkg-config (1.4.6)
|
21
21
|
rake (13.0.6)
|
data/README.md
CHANGED
@@ -1,15 +1,31 @@
|
|
1
1
|
# Parqueteur
|
2
2
|
|
3
|
-
|
3
|
+
[![Gem Version](https://badge.fury.io/rb/parqueteur.svg)](https://badge.fury.io/rb/parqueteur)
|
4
4
|
|
5
|
-
|
5
|
+
Parqueteur enables you to generate Apache Parquet files from raw data.
|
6
6
|
|
7
|
+
## Dependencies
|
8
|
+
|
9
|
+
Since I only tested Parqueteur on Ubuntu, I don't have any install scripts for others operating systems.
|
10
|
+
### Debian/Ubuntu packages
|
11
|
+
- `libgirepository1.0-dev`
|
12
|
+
- `libarrow-dev`
|
13
|
+
- `libarrow-glib-dev`
|
14
|
+
- `libparquet-dev`
|
15
|
+
- `libparquet-glib-dev`
|
16
|
+
|
17
|
+
You can check `scripts/apache-arrow-ubuntu-install.sh` script for a quick way to install all of them.
|
7
18
|
## Installation
|
8
19
|
|
9
20
|
Add this line to your application's Gemfile:
|
10
21
|
|
11
22
|
```ruby
|
12
|
-
gem 'parqueteur'
|
23
|
+
gem 'parqueteur', '~> 1.0'
|
24
|
+
```
|
25
|
+
|
26
|
+
> (optional) If you don't want to require Parqueteur globally you can add `require: false` to the Gemfile instruction:
|
27
|
+
```ruby
|
28
|
+
gem 'parqueteur', '~> 1.0', require: false
|
13
29
|
```
|
14
30
|
|
15
31
|
And then execute:
|
@@ -22,14 +38,127 @@ Or install it yourself as:
|
|
22
38
|
|
23
39
|
## Usage
|
24
40
|
|
25
|
-
|
41
|
+
Parqueteur provides an elegant way to generate Apache Parquet files from a defined schema.
|
42
|
+
|
43
|
+
Converters accepts any object that implements `Enumerable` as data source.
|
44
|
+
|
45
|
+
### Working example
|
46
|
+
|
47
|
+
```ruby
|
48
|
+
require 'parqueteur'
|
49
|
+
|
50
|
+
class FooParquetConverter < Parqueteur::Converter
|
51
|
+
column :id, :bigint
|
52
|
+
column :reference, :string
|
53
|
+
column :datetime, :timestamp
|
54
|
+
end
|
55
|
+
|
56
|
+
data = [
|
57
|
+
{ 'id' => 1, 'reference' => 'hello world 1', 'datetime' => Time.now },
|
58
|
+
{ 'id' => 2, 'reference' => 'hello world 2', 'datetime' => Time.now },
|
59
|
+
{ 'id' => 3, 'reference' => 'hello world 3', 'datetime' => Time.now }
|
60
|
+
]
|
61
|
+
|
62
|
+
# initialize Converter with Parquet GZIP compression mode
|
63
|
+
converter = FooParquetConverter.new(data, compression: :gzip)
|
26
64
|
|
27
|
-
|
65
|
+
# write result to file
|
66
|
+
converter.write('hello_world.parquet')
|
28
67
|
|
29
|
-
|
68
|
+
# in-memory result (StringIO)
|
69
|
+
converter.to_io
|
70
|
+
|
71
|
+
# write to temporary file (Tempfile)
|
72
|
+
# don't forget to `close` / `unlink` it after usage
|
73
|
+
converter.to_tmpfile
|
74
|
+
|
75
|
+
# convert to Arrow::Table
|
76
|
+
pp converter.to_arrow_table
|
77
|
+
```
|
78
|
+
|
79
|
+
### Using transformers
|
80
|
+
|
81
|
+
You can use transformers to apply data items transformations.
|
82
|
+
|
83
|
+
From `examples/cars.rb`:
|
84
|
+
|
85
|
+
```ruby
|
86
|
+
require 'parqueteur'
|
87
|
+
|
88
|
+
class Car
|
89
|
+
attr_reader :name, :production_year
|
90
|
+
|
91
|
+
def initialize(name, production_year)
|
92
|
+
@name = name
|
93
|
+
@production_year = production_year
|
94
|
+
end
|
95
|
+
end
|
96
|
+
|
97
|
+
class CarParquetConverter < Parqueteur::Converter
|
98
|
+
column :name, :string
|
99
|
+
column :production_year, :integer
|
100
|
+
|
101
|
+
transform do |car|
|
102
|
+
{
|
103
|
+
'name' => car.name,
|
104
|
+
'production_year' => car.production_year
|
105
|
+
}
|
106
|
+
end
|
107
|
+
end
|
108
|
+
|
109
|
+
cars = [
|
110
|
+
Car.new('Alfa Romeo 75', 1985),
|
111
|
+
Car.new('Alfa Romeo 33', 1983),
|
112
|
+
Car.new('Audi A3', 1996),
|
113
|
+
Car.new('Audi A4', 1994),
|
114
|
+
Car.new('BMW 503', 1956),
|
115
|
+
Car.new('BMW X5', 1999)
|
116
|
+
]
|
117
|
+
|
118
|
+
# initialize Converter with Parquet GZIP compression mode
|
119
|
+
converter = CarParquetConverter.new(data, compression: :gzip)
|
120
|
+
|
121
|
+
# write result to file
|
122
|
+
pp converter.to_arrow_table
|
123
|
+
```
|
124
|
+
|
125
|
+
Output:
|
126
|
+
```
|
127
|
+
#<Arrow::Table:0x7fc1fb24b958 ptr=0x7fc1faedd910>
|
128
|
+
# name production_year
|
129
|
+
0 Alfa Romeo 75 1985
|
130
|
+
1 Alfa Romeo 33 1983
|
131
|
+
2 Audi A3 1996
|
132
|
+
3 Audi A4 1994
|
133
|
+
4 BMW 503 1956
|
134
|
+
5 BMW X5 1999
|
135
|
+
```
|
30
136
|
|
31
|
-
|
137
|
+
### Available Types
|
138
|
+
|
139
|
+
| Name (Symbol) | Apache Parquet Type |
|
140
|
+
| ------------- | --------- |
|
141
|
+
| `:array` | `Array` |
|
142
|
+
| `:bigdecimal` | `Decimal256` |
|
143
|
+
| `:bigint` | `Int64` or `UInt64` with `unsigned: true` option |
|
144
|
+
| `:boolean` | `Boolean` |
|
145
|
+
| `:date` | `Date32` |
|
146
|
+
| `:date32` | `Date32` |
|
147
|
+
| `:date64` | `Date64` |
|
148
|
+
| `:decimal` | `Decimal128` |
|
149
|
+
| `:decimal128` | `Decimal128` |
|
150
|
+
| `:decimal256` | `Decimal256` |
|
151
|
+
| `:int32` | `Int32` or `UInt32` with `unsigned: true` option |
|
152
|
+
| `:int64` | `Int64` or `UInt64` with `unsigned: true` option |
|
153
|
+
| `:integer` | `Int32` or `UInt32` with `unsigned: true` option |
|
154
|
+
| `:map` | `Map` |
|
155
|
+
| `:string` | `String` |
|
156
|
+
| `:struct` | `Struct` |
|
157
|
+
| `:time` | `Time32` |
|
158
|
+
| `:time32` | `Time32` |
|
159
|
+
| `:time64` | `Time64` |
|
160
|
+
| `:timestamp` | `Timestamp` |
|
32
161
|
|
33
162
|
## Contributing
|
34
163
|
|
35
|
-
Bug reports and pull requests are welcome on GitHub at https://github.com/
|
164
|
+
Bug reports and pull requests are welcome on GitHub at https://github.com/pocketsizesun/parqueteur-ruby.
|
data/examples/cars.rb
ADDED
@@ -0,0 +1,40 @@
|
|
1
|
+
require 'bundler/setup'
|
2
|
+
require 'parqueteur'
|
3
|
+
|
4
|
+
class Car
|
5
|
+
attr_reader :name, :production_year
|
6
|
+
|
7
|
+
def initialize(name, production_year)
|
8
|
+
@name = name
|
9
|
+
@production_year = production_year
|
10
|
+
end
|
11
|
+
end
|
12
|
+
|
13
|
+
class CarParquetConverter < Parqueteur::Converter
|
14
|
+
column :name, :string
|
15
|
+
column :production_year, :integer
|
16
|
+
|
17
|
+
transform do |car|
|
18
|
+
{
|
19
|
+
'name' => car.name,
|
20
|
+
'production_year' => car.production_year
|
21
|
+
}
|
22
|
+
end
|
23
|
+
end
|
24
|
+
|
25
|
+
cars = [
|
26
|
+
Car.new('Alfa Romeo 75', 1985),
|
27
|
+
Car.new('Alfa Romeo 33', 1983),
|
28
|
+
Car.new('Audi A3', 1996),
|
29
|
+
Car.new('Audi A4', 1994),
|
30
|
+
Car.new('BMW 503', 1956),
|
31
|
+
Car.new('BMW X5', 1999)
|
32
|
+
]
|
33
|
+
|
34
|
+
# initialize Converter with Parquet GZIP compression mode
|
35
|
+
converter = CarParquetConverter.new(
|
36
|
+
cars, compression: :gzip
|
37
|
+
)
|
38
|
+
|
39
|
+
# write result to file
|
40
|
+
pp converter.to_arrow_table
|
@@ -0,0 +1,56 @@
|
|
1
|
+
require 'bundler/setup'
|
2
|
+
require 'parqueteur'
|
3
|
+
require 'securerandom'
|
4
|
+
require 'benchmark'
|
5
|
+
|
6
|
+
class FooParquetConverter < Parqueteur::Converter
|
7
|
+
column :id, :bigint
|
8
|
+
column :reference, :string
|
9
|
+
column :hash, :map, key: :string, value: :string
|
10
|
+
# column :hash2, :map, key: :string, value: :string
|
11
|
+
# column :hash3, :map, key: :string, value: :string
|
12
|
+
column :valid, :boolean
|
13
|
+
column :total, :integer
|
14
|
+
column :numbers, :array, elements: :integer
|
15
|
+
column :my_struct, :struct do
|
16
|
+
field :test, :string
|
17
|
+
field :mon_nombre, :integer
|
18
|
+
end
|
19
|
+
end
|
20
|
+
|
21
|
+
def random_hash
|
22
|
+
{
|
23
|
+
'a' => SecureRandom.hex(128),
|
24
|
+
'b' => SecureRandom.hex(128),
|
25
|
+
'c' => SecureRandom.hex(128),
|
26
|
+
'd' => SecureRandom.hex(128),
|
27
|
+
'e' => SecureRandom.hex(128),
|
28
|
+
'f' => SecureRandom.hex(128),
|
29
|
+
'g' => SecureRandom.hex(128),
|
30
|
+
'h' => SecureRandom.hex(128),
|
31
|
+
'i' => SecureRandom.hex(128),
|
32
|
+
'j' => SecureRandom.hex(128),
|
33
|
+
'k' => SecureRandom.hex(128),
|
34
|
+
}
|
35
|
+
end
|
36
|
+
|
37
|
+
data = 10000.times.collect do |i|
|
38
|
+
{
|
39
|
+
'id' => i + 1,
|
40
|
+
'reference' => "coucou:#{i}",
|
41
|
+
'hash' => random_hash,
|
42
|
+
# 'hash2' => random_hash,
|
43
|
+
# 'hash3' => random_hash,
|
44
|
+
'valid' => rand < 0.5,
|
45
|
+
'total' => rand(100..500),
|
46
|
+
'numbers' => [1, 2, 3]
|
47
|
+
}
|
48
|
+
end
|
49
|
+
puts "data generation OK"
|
50
|
+
|
51
|
+
# initialize Converter with Parquet GZIP compression mode
|
52
|
+
converter = FooParquetConverter.new(data, compression: :gzip)
|
53
|
+
|
54
|
+
# write result to file
|
55
|
+
converter.write('tmp/example.gzip-compressed.parquet')
|
56
|
+
converter.write('tmp/example.no-gzip.parquet', compression: false)
|
@@ -0,0 +1,54 @@
|
|
1
|
+
require 'bundler/setup'
|
2
|
+
require 'parqueteur'
|
3
|
+
require 'securerandom'
|
4
|
+
require 'benchmark'
|
5
|
+
|
6
|
+
class Foo < Parqueteur::Converter
|
7
|
+
column :id, :bigint
|
8
|
+
column :reference, :string
|
9
|
+
column :hash, :map, key: :string, value: :string
|
10
|
+
# column :hash2, :map, key: :string, value: :string
|
11
|
+
# column :hash3, :map, key: :string, value: :string
|
12
|
+
column :valid, :boolean
|
13
|
+
column :total, :integer
|
14
|
+
column :numbers, :array, elements: :integer
|
15
|
+
column :my_struct, :struct do
|
16
|
+
field :test, :string
|
17
|
+
field :mon_nombre, :integer
|
18
|
+
end
|
19
|
+
end
|
20
|
+
|
21
|
+
def random_hash
|
22
|
+
{
|
23
|
+
'a' => SecureRandom.hex(128),
|
24
|
+
'b' => SecureRandom.hex(128),
|
25
|
+
'c' => SecureRandom.hex(128),
|
26
|
+
'd' => SecureRandom.hex(128),
|
27
|
+
'e' => SecureRandom.hex(128),
|
28
|
+
'f' => SecureRandom.hex(128),
|
29
|
+
'g' => SecureRandom.hex(128),
|
30
|
+
'h' => SecureRandom.hex(128),
|
31
|
+
'i' => SecureRandom.hex(128),
|
32
|
+
'j' => SecureRandom.hex(128),
|
33
|
+
'k' => SecureRandom.hex(128),
|
34
|
+
}
|
35
|
+
end
|
36
|
+
|
37
|
+
data = 10000.times.collect do |i|
|
38
|
+
{
|
39
|
+
'id' => i + 1,
|
40
|
+
'reference' => "coucou:#{i}",
|
41
|
+
'hash' => random_hash,
|
42
|
+
# 'hash2' => random_hash,
|
43
|
+
# 'hash3' => random_hash,
|
44
|
+
'valid' => rand < 0.5,
|
45
|
+
'total' => rand(100..500),
|
46
|
+
'numbers' => [1, 2, 3]
|
47
|
+
}
|
48
|
+
end
|
49
|
+
puts "data generation OK"
|
50
|
+
|
51
|
+
converter = Foo.new(data, compression: :gzip)
|
52
|
+
pp converter.to_io
|
53
|
+
pp converter.to_arrow_table
|
54
|
+
converter.write('tmp/test.parquet')
|
@@ -0,0 +1,52 @@
|
|
1
|
+
require 'bundler/setup'
|
2
|
+
require 'parqueteur'
|
3
|
+
require 'securerandom'
|
4
|
+
require 'benchmark'
|
5
|
+
|
6
|
+
class Foo < Parqueteur::Converter
|
7
|
+
column :id, :bigint
|
8
|
+
column :reference, :string
|
9
|
+
column :hash, :map, key: :string, value: :string
|
10
|
+
# column :hash2, :map, key: :string, value: :string
|
11
|
+
# column :hash3, :map, key: :string, value: :string
|
12
|
+
column :valid, :boolean
|
13
|
+
column :total, :integer
|
14
|
+
column :numbers, :array, elements: :integer
|
15
|
+
column :my_struct, :struct do
|
16
|
+
field :test, :string
|
17
|
+
field :mon_nombre, :integer
|
18
|
+
end
|
19
|
+
end
|
20
|
+
|
21
|
+
def random_hash
|
22
|
+
{
|
23
|
+
'a' => SecureRandom.hex(128),
|
24
|
+
'b' => SecureRandom.hex(128),
|
25
|
+
'c' => SecureRandom.hex(128),
|
26
|
+
'd' => SecureRandom.hex(128),
|
27
|
+
'e' => SecureRandom.hex(128),
|
28
|
+
'f' => SecureRandom.hex(128),
|
29
|
+
'g' => SecureRandom.hex(128),
|
30
|
+
'h' => SecureRandom.hex(128),
|
31
|
+
'i' => SecureRandom.hex(128),
|
32
|
+
'j' => SecureRandom.hex(128),
|
33
|
+
'k' => SecureRandom.hex(128),
|
34
|
+
}
|
35
|
+
end
|
36
|
+
|
37
|
+
data = 10000.times.collect do |i|
|
38
|
+
{
|
39
|
+
'id' => i + 1,
|
40
|
+
'reference' => "coucou:#{i}",
|
41
|
+
'hash' => random_hash,
|
42
|
+
# 'hash2' => random_hash,
|
43
|
+
# 'hash3' => random_hash,
|
44
|
+
'valid' => rand < 0.5,
|
45
|
+
'total' => rand(100..500),
|
46
|
+
'numbers' => [1, 2, 3]
|
47
|
+
}
|
48
|
+
end
|
49
|
+
puts "data generation OK"
|
50
|
+
|
51
|
+
io = Foo.convert(data)
|
52
|
+
pp io.read
|
@@ -0,0 +1,54 @@
|
|
1
|
+
require 'bundler/setup'
|
2
|
+
require 'parqueteur'
|
3
|
+
require 'securerandom'
|
4
|
+
require 'benchmark'
|
5
|
+
|
6
|
+
class Foo < Parqueteur::Converter
|
7
|
+
column :id, :bigint
|
8
|
+
column :reference, :string
|
9
|
+
column :hash, :map, key: :string, value: :string
|
10
|
+
# column :hash2, :map, key: :string, value: :string
|
11
|
+
# column :hash3, :map, key: :string, value: :string
|
12
|
+
column :valid, :boolean
|
13
|
+
column :total, :integer
|
14
|
+
column :numbers, :array, elements: :integer
|
15
|
+
column :my_struct, :struct do
|
16
|
+
field :test, :string
|
17
|
+
field :mon_nombre, :integer
|
18
|
+
end
|
19
|
+
end
|
20
|
+
|
21
|
+
def random_hash
|
22
|
+
{
|
23
|
+
'a' => SecureRandom.hex(128),
|
24
|
+
'b' => SecureRandom.hex(128),
|
25
|
+
'c' => SecureRandom.hex(128),
|
26
|
+
'd' => SecureRandom.hex(128),
|
27
|
+
'e' => SecureRandom.hex(128),
|
28
|
+
'f' => SecureRandom.hex(128),
|
29
|
+
'g' => SecureRandom.hex(128),
|
30
|
+
'h' => SecureRandom.hex(128),
|
31
|
+
'i' => SecureRandom.hex(128),
|
32
|
+
'j' => SecureRandom.hex(128),
|
33
|
+
'k' => SecureRandom.hex(128),
|
34
|
+
}
|
35
|
+
end
|
36
|
+
|
37
|
+
data = 10000.times.collect do |i|
|
38
|
+
{
|
39
|
+
'id' => i + 1,
|
40
|
+
'reference' => "coucou:#{i}",
|
41
|
+
'hash' => random_hash,
|
42
|
+
# 'hash2' => random_hash,
|
43
|
+
# 'hash3' => random_hash,
|
44
|
+
'valid' => rand < 0.5,
|
45
|
+
'total' => rand(100..500),
|
46
|
+
'numbers' => [1, 2, 3]
|
47
|
+
}
|
48
|
+
end
|
49
|
+
puts "data generation OK"
|
50
|
+
|
51
|
+
converter = Foo.new(data, compression: :gzip)
|
52
|
+
converter.split(200).each_with_index do |chunk, idx|
|
53
|
+
puts "#{idx}: #{chunk.path}"
|
54
|
+
end
|
@@ -0,0 +1,52 @@
|
|
1
|
+
require 'bundler/setup'
|
2
|
+
require 'parqueteur'
|
3
|
+
require 'securerandom'
|
4
|
+
require 'benchmark'
|
5
|
+
|
6
|
+
class Foo < Parqueteur::Converter
|
7
|
+
column :id, :bigint
|
8
|
+
column :reference, :string
|
9
|
+
column :hash, :map, key: :string, value: :string
|
10
|
+
# column :hash2, :map, key: :string, value: :string
|
11
|
+
# column :hash3, :map, key: :string, value: :string
|
12
|
+
column :valid, :boolean
|
13
|
+
column :total, :integer
|
14
|
+
column :numbers, :array, elements: :integer
|
15
|
+
column :my_struct, :struct do
|
16
|
+
field :test, :string
|
17
|
+
field :mon_nombre, :integer
|
18
|
+
end
|
19
|
+
end
|
20
|
+
|
21
|
+
def random_hash
|
22
|
+
{
|
23
|
+
'a' => SecureRandom.hex(128),
|
24
|
+
'b' => SecureRandom.hex(128),
|
25
|
+
'c' => SecureRandom.hex(128),
|
26
|
+
'd' => SecureRandom.hex(128),
|
27
|
+
'e' => SecureRandom.hex(128),
|
28
|
+
'f' => SecureRandom.hex(128),
|
29
|
+
'g' => SecureRandom.hex(128),
|
30
|
+
'h' => SecureRandom.hex(128),
|
31
|
+
'i' => SecureRandom.hex(128),
|
32
|
+
'j' => SecureRandom.hex(128),
|
33
|
+
'k' => SecureRandom.hex(128),
|
34
|
+
}
|
35
|
+
end
|
36
|
+
|
37
|
+
data = 10000.times.collect do |i|
|
38
|
+
{
|
39
|
+
'id' => i + 1,
|
40
|
+
'reference' => "coucou:#{i}",
|
41
|
+
'hash' => random_hash,
|
42
|
+
# 'hash2' => random_hash,
|
43
|
+
# 'hash3' => random_hash,
|
44
|
+
'valid' => rand < 0.5,
|
45
|
+
'total' => rand(100..500),
|
46
|
+
'numbers' => [1, 2, 3]
|
47
|
+
}
|
48
|
+
end
|
49
|
+
puts "data generation OK"
|
50
|
+
|
51
|
+
path = 'tmp/test.parquet'
|
52
|
+
Foo.convert_to(data, path)
|
@@ -0,0 +1,57 @@
|
|
1
|
+
require 'bundler/setup'
|
2
|
+
require 'parqueteur'
|
3
|
+
|
4
|
+
class FooParquetConverter < Parqueteur::Converter
|
5
|
+
column :id, :bigint
|
6
|
+
column :my_string_array, :array, elements: :string
|
7
|
+
column :my_date, :date
|
8
|
+
column :my_decimal, :decimal, precision: 12, scale: 4
|
9
|
+
column :my_int, :integer
|
10
|
+
column :my_map, :map, key: :string, value: :string
|
11
|
+
column :my_string, :string
|
12
|
+
column :my_struct, :struct do
|
13
|
+
field :my_struct_str, :string
|
14
|
+
field :my_struct_int, :integer
|
15
|
+
end
|
16
|
+
column :my_time, :time
|
17
|
+
column :my_timestamp, :timestamp
|
18
|
+
end
|
19
|
+
|
20
|
+
data = 100.times.collect do |i|
|
21
|
+
{
|
22
|
+
'id' => i,
|
23
|
+
'my_string_array' => %w[a b c],
|
24
|
+
'my_date' => Date.today,
|
25
|
+
'my_decimal' => BigDecimal('0.03'),
|
26
|
+
'my_int' => rand(1..10),
|
27
|
+
'my_map' => { 'a' => 'b' },
|
28
|
+
'my_string' => 'Hello World',
|
29
|
+
'my_struct' => {
|
30
|
+
'my_struct_str' => 'Hello World',
|
31
|
+
'my_struct_int' => 1
|
32
|
+
},
|
33
|
+
'my_time' => 3600,
|
34
|
+
'my_timestamp' => Time.now
|
35
|
+
}
|
36
|
+
end
|
37
|
+
|
38
|
+
# initialize Converter with Parquet GZIP compression mode
|
39
|
+
converter = FooParquetConverter.new(data, compression: :gzip)
|
40
|
+
|
41
|
+
# write result to file
|
42
|
+
converter.write('tmp/hello_world.compressed.parquet')
|
43
|
+
converter.write('tmp/hello_world.parquet', compression: false)
|
44
|
+
|
45
|
+
# in-memory result (StringIO)
|
46
|
+
converter.to_io
|
47
|
+
|
48
|
+
# write to temporary file (Tempfile)
|
49
|
+
# don't forget to `close` / `unlink` it after usage
|
50
|
+
converter.to_tmpfile
|
51
|
+
|
52
|
+
# Arrow Table
|
53
|
+
table = converter.to_arrow_table
|
54
|
+
table.each_record do |record|
|
55
|
+
# pp record['my_decimal'].to_f
|
56
|
+
pp record.to_h
|
57
|
+
end
|
@@ -0,0 +1,44 @@
|
|
1
|
+
require 'bundler/setup'
|
2
|
+
require 'parqueteur'
|
3
|
+
|
4
|
+
class FooParquetConverter < Parqueteur::Converter
|
5
|
+
column :id, :bigint
|
6
|
+
column :reference, :string
|
7
|
+
column :datetime, :timestamp
|
8
|
+
column :beers_count, :integer
|
9
|
+
|
10
|
+
transform do |item|
|
11
|
+
item.merge(
|
12
|
+
'datetime' => Time.now
|
13
|
+
)
|
14
|
+
end
|
15
|
+
|
16
|
+
transform :add_beers
|
17
|
+
|
18
|
+
private
|
19
|
+
|
20
|
+
def add_beers(item)
|
21
|
+
item['beers_count'] += rand(1..3)
|
22
|
+
item
|
23
|
+
end
|
24
|
+
end
|
25
|
+
|
26
|
+
data = 10.times.lazy.map do |i|
|
27
|
+
{ 'id' => i + 1, 'reference' => 'hello world 1', 'beers_count' => 0 }
|
28
|
+
end
|
29
|
+
|
30
|
+
# initialize Converter with Parquet GZIP compression mode
|
31
|
+
converter = FooParquetConverter.new(data, compression: :gzip)
|
32
|
+
|
33
|
+
# write result to file
|
34
|
+
converter.write('tmp/hello_world.parquet')
|
35
|
+
|
36
|
+
# in-memory result (StringIO)
|
37
|
+
converter.to_io
|
38
|
+
|
39
|
+
# write to temporary file (Tempfile)
|
40
|
+
# don't forget to `close` / `unlink` it after usage
|
41
|
+
converter.to_tmpfile
|
42
|
+
|
43
|
+
# convert to Arrow::Table
|
44
|
+
pp converter.to_arrow_table
|
data/lib/parqueteur/converter.rb
CHANGED
@@ -2,7 +2,7 @@
|
|
2
2
|
|
3
3
|
module Parqueteur
|
4
4
|
class Converter
|
5
|
-
|
5
|
+
DEFAULT_BATCH_SIZE = 10
|
6
6
|
|
7
7
|
def self.inline(&block)
|
8
8
|
Class.new(self, &block)
|
@@ -24,104 +24,137 @@ module Parqueteur
|
|
24
24
|
transforms << (method_name || block)
|
25
25
|
end
|
26
26
|
|
27
|
-
def self.convert(input,
|
28
|
-
|
29
|
-
|
30
|
-
|
31
|
-
|
32
|
-
|
27
|
+
def self.convert(input, **kwargs)
|
28
|
+
new(input, **kwargs).to_io
|
29
|
+
end
|
30
|
+
|
31
|
+
def self.convert_to(input, output_path, **kwargs)
|
32
|
+
converter = new(input, **kwargs)
|
33
|
+
converter.write(output_path)
|
34
|
+
end
|
35
|
+
|
36
|
+
# @param [Enumerable] An enumerable object
|
37
|
+
# @option [Symbol] compression - :gzip
|
38
|
+
def initialize(input, **kwargs)
|
39
|
+
@input = Parqueteur::Input.from(input)
|
40
|
+
@batch_size = kwargs.fetch(:batch_size, DEFAULT_BATCH_SIZE)
|
41
|
+
@compression = kwargs.fetch(:compression, nil)&.to_sym
|
42
|
+
end
|
43
|
+
|
44
|
+
def split(size, batch_size: nil, compression: nil)
|
45
|
+
Enumerator.new do |arr|
|
46
|
+
options = {
|
47
|
+
batch_size: batch_size || @batch_size,
|
48
|
+
compression: compression || @compression
|
49
|
+
}
|
50
|
+
@input.each_slice(size) do |records|
|
51
|
+
local_converter = self.class.new(records, **options)
|
52
|
+
file = local_converter.to_tmpfile
|
53
|
+
arr << file
|
54
|
+
file.close
|
55
|
+
file.unlink
|
56
|
+
end
|
33
57
|
end
|
34
58
|
end
|
35
59
|
|
36
|
-
def
|
37
|
-
|
38
|
-
|
39
|
-
|
40
|
-
|
41
|
-
|
42
|
-
|
43
|
-
|
44
|
-
|
45
|
-
|
46
|
-
|
47
|
-
|
48
|
-
|
49
|
-
|
50
|
-
|
51
|
-
|
52
|
-
|
53
|
-
|
60
|
+
def split_by_io(size, batch_size: nil, compression: nil)
|
61
|
+
Enumerator.new do |arr|
|
62
|
+
options = {
|
63
|
+
batch_size: batch_size || @batch_size,
|
64
|
+
compression: compression || @compression
|
65
|
+
}
|
66
|
+
@input.each_slice(size) do |records|
|
67
|
+
local_converter = self.class.new(records, **options)
|
68
|
+
arr << local_converter.to_io
|
69
|
+
end
|
70
|
+
end
|
71
|
+
end
|
72
|
+
|
73
|
+
def write(path, batch_size: nil, compression: nil)
|
74
|
+
compression = @compression if compression.nil?
|
75
|
+
batch_size = @batch_size if batch_size.nil?
|
76
|
+
arrow_schema = self.class.columns.arrow_schema
|
77
|
+
writer_properties = Parquet::WriterProperties.new
|
78
|
+
if !compression.nil? && compression != false
|
79
|
+
writer_properties.set_compression(compression)
|
80
|
+
end
|
81
|
+
|
82
|
+
Arrow::FileOutputStream.open(path, false) do |output|
|
83
|
+
Parquet::ArrowFileWriter.open(arrow_schema, output, writer_properties) do |writer|
|
84
|
+
@input.each_slice(batch_size) do |records|
|
85
|
+
arrow_table = build_arrow_table(records)
|
86
|
+
writer.write_table(arrow_table, 1024)
|
87
|
+
end
|
88
|
+
end
|
54
89
|
end
|
90
|
+
|
91
|
+
true
|
92
|
+
end
|
93
|
+
|
94
|
+
def to_tmpfile(options = {})
|
95
|
+
tempfile = Tempfile.new
|
96
|
+
tempfile.binmode
|
97
|
+
write(tempfile.path, **options)
|
98
|
+
tempfile.rewind
|
99
|
+
tempfile
|
55
100
|
end
|
56
101
|
|
57
|
-
def
|
58
|
-
|
102
|
+
def to_io(options = {})
|
103
|
+
tmpfile = to_tmpfile(options)
|
104
|
+
strio = StringIO.new(tmpfile.read)
|
105
|
+
tmpfile.close
|
106
|
+
tmpfile.unlink
|
107
|
+
strio
|
59
108
|
end
|
60
109
|
|
61
|
-
def
|
62
|
-
|
110
|
+
def to_arrow_table(options = {})
|
111
|
+
file = to_tmpfile(options)
|
112
|
+
table = Arrow::Table.load(file.path, format: :parquet)
|
113
|
+
file.close
|
114
|
+
file.unlink
|
115
|
+
table
|
63
116
|
end
|
64
117
|
|
65
|
-
def to_blob
|
66
|
-
|
118
|
+
def to_blob(options = {})
|
119
|
+
to_tmpfile(options).read
|
67
120
|
end
|
68
121
|
|
69
|
-
|
122
|
+
private
|
123
|
+
|
124
|
+
def build_arrow_table(records)
|
70
125
|
transforms = self.class.transforms
|
71
126
|
|
72
|
-
|
127
|
+
values = self.class.columns.each_with_object({}) do |column, hash|
|
73
128
|
hash[column.name] = []
|
74
129
|
end
|
75
|
-
items_count = 0
|
76
|
-
@input.each_slice(100) do |items|
|
77
|
-
values = self.class.columns.each_with_object({}) do |column, hash|
|
78
|
-
hash[column.name] = []
|
79
|
-
end
|
80
130
|
|
81
|
-
|
82
|
-
|
83
|
-
|
84
|
-
|
85
|
-
|
86
|
-
|
87
|
-
|
88
|
-
|
89
|
-
|
90
|
-
end
|
131
|
+
records.each do |item|
|
132
|
+
if transforms.length > 0
|
133
|
+
transforms.each do |transform|
|
134
|
+
item = \
|
135
|
+
if transform.is_a?(Symbol)
|
136
|
+
__send__(transform, item)
|
137
|
+
else
|
138
|
+
transform.call(item)
|
139
|
+
end
|
91
140
|
end
|
141
|
+
end
|
92
142
|
|
93
|
-
|
94
|
-
|
95
|
-
|
96
|
-
|
97
|
-
|
98
|
-
end
|
143
|
+
values.each_key do |value_key|
|
144
|
+
if item.key?(value_key)
|
145
|
+
values[value_key] << item[value_key]
|
146
|
+
else
|
147
|
+
values[value_key] << nil
|
99
148
|
end
|
100
149
|
end
|
150
|
+
end
|
101
151
|
|
102
|
-
|
152
|
+
Arrow::Table.new(
|
153
|
+
values.each_with_object({}) do |item, hash|
|
103
154
|
column = self.class.columns.find(item[0])
|
104
|
-
hash[item[0]].
|
105
|
-
column.type.build_value_array(item[1])
|
106
|
-
)
|
155
|
+
hash[item[0]] = column.type.build_value_array(item[1])
|
107
156
|
end
|
108
|
-
|
109
|
-
items_count += items.length
|
110
|
-
end
|
111
|
-
|
112
|
-
if items_count > 0
|
113
|
-
Arrow::Table.new(
|
114
|
-
chunks.transform_values! do |value|
|
115
|
-
Arrow::ChunkedArray.new(value)
|
116
|
-
end
|
117
|
-
)
|
118
|
-
else
|
119
|
-
Arrow::Table.new(
|
120
|
-
self.class.columns.each_with_object({}) do |column, hash|
|
121
|
-
hash[column.name] = column.type.build_value_array([])
|
122
|
-
end
|
123
|
-
)
|
124
|
-
end
|
157
|
+
)
|
125
158
|
end
|
126
159
|
end
|
127
160
|
end
|
data/lib/parqueteur/input.rb
CHANGED
@@ -4,40 +4,25 @@ module Parqueteur
|
|
4
4
|
class Input
|
5
5
|
include Enumerable
|
6
6
|
|
7
|
-
def self.from(arg
|
8
|
-
|
9
|
-
|
10
|
-
|
11
|
-
if File.exist?(arg)
|
12
|
-
File.new(arg, 'r')
|
13
|
-
else
|
14
|
-
arg.split("\n")
|
15
|
-
end
|
16
|
-
when Enumerable
|
17
|
-
arg
|
18
|
-
end,
|
19
|
-
options
|
20
|
-
)
|
7
|
+
def self.from(arg)
|
8
|
+
return arg if arg.is_a?(self)
|
9
|
+
|
10
|
+
new(arg)
|
21
11
|
end
|
22
12
|
|
23
|
-
def initialize(source
|
13
|
+
def initialize(source)
|
14
|
+
unless source.is_a?(Enumerable)
|
15
|
+
raise ArgumentError, 'Enumerable object expected'
|
16
|
+
end
|
17
|
+
|
24
18
|
@source = source
|
25
|
-
@options = options
|
26
19
|
end
|
27
20
|
|
28
21
|
def each(&block)
|
29
|
-
|
30
|
-
when File
|
31
|
-
if @options.fetch(:json_newlines, true) == true
|
32
|
-
@source.each_line do |line|
|
33
|
-
yield(JSON.parse(line.strip))
|
34
|
-
end
|
35
|
-
else
|
36
|
-
JSON.parse(@source.read).each(&block)
|
37
|
-
end
|
38
|
-
@source.rewind
|
39
|
-
when Enumerable
|
22
|
+
if block_given?
|
40
23
|
@source.each(&block)
|
24
|
+
else
|
25
|
+
@source.to_enum(:each)
|
41
26
|
end
|
42
27
|
end
|
43
28
|
end
|
@@ -7,14 +7,24 @@ module Parqueteur
|
|
7
7
|
def self.registered_types
|
8
8
|
@registered_types ||= {
|
9
9
|
array: Parqueteur::Types::ArrayType,
|
10
|
+
bigdecimal: Parqueteur::Types::Decimal256Type,
|
10
11
|
bigint: Parqueteur::Types::Int64Type,
|
11
12
|
boolean: Parqueteur::Types::BooleanType,
|
13
|
+
date: Parqueteur::Types::Date32Type,
|
14
|
+
date32: Parqueteur::Types::Date32Type,
|
15
|
+
date64: Parqueteur::Types::Date64Type,
|
16
|
+
decimal: Parqueteur::Types::Decimal128Type,
|
17
|
+
decimal128: Parqueteur::Types::Decimal128Type,
|
18
|
+
decimal256: Parqueteur::Types::Decimal256Type,
|
12
19
|
int32: Parqueteur::Types::Int32Type,
|
13
20
|
int64: Parqueteur::Types::Int64Type,
|
14
21
|
integer: Parqueteur::Types::Int32Type,
|
15
22
|
map: Parqueteur::Types::MapType,
|
16
23
|
string: Parqueteur::Types::StringType,
|
17
24
|
struct: Parqueteur::Types::StructType,
|
25
|
+
time: Parqueteur::Types::Time32Type,
|
26
|
+
time32: Parqueteur::Types::Time32Type,
|
27
|
+
time64: Parqueteur::Types::Time64Type,
|
18
28
|
timestamp: Parqueteur::Types::TimestampType
|
19
29
|
}
|
20
30
|
end
|
@@ -0,0 +1,15 @@
|
|
1
|
+
# frozen_string_literal: true
|
2
|
+
|
3
|
+
module Parqueteur
|
4
|
+
module Types
|
5
|
+
class Date32Type < Parqueteur::Type
|
6
|
+
def build_value_array(values)
|
7
|
+
Arrow::Date32ArrayBuilder.build(values)
|
8
|
+
end
|
9
|
+
|
10
|
+
def arrow_type_builder
|
11
|
+
Arrow::Date32DataType.new
|
12
|
+
end
|
13
|
+
end
|
14
|
+
end
|
15
|
+
end
|
@@ -0,0 +1,15 @@
|
|
1
|
+
# frozen_string_literal: true
|
2
|
+
|
3
|
+
module Parqueteur
|
4
|
+
module Types
|
5
|
+
class Date64Type < Parqueteur::Type
|
6
|
+
def build_value_array(values)
|
7
|
+
Arrow::Date64ArrayBuilder.build([values])
|
8
|
+
end
|
9
|
+
|
10
|
+
def arrow_type_builder
|
11
|
+
Arrow::Date64DataType.new
|
12
|
+
end
|
13
|
+
end
|
14
|
+
end
|
15
|
+
end
|
@@ -0,0 +1,29 @@
|
|
1
|
+
# frozen_string_literal: true
|
2
|
+
|
3
|
+
module Parqueteur
|
4
|
+
module Types
|
5
|
+
class Decimal128Type < Parqueteur::Type
|
6
|
+
def initialize(options = {}, &block)
|
7
|
+
@scale = options.fetch(:scale)
|
8
|
+
@precision = options.fetch(:precision)
|
9
|
+
@format_str = "%.#{@scale}f"
|
10
|
+
super(options, &block)
|
11
|
+
end
|
12
|
+
|
13
|
+
def build_value_array(values)
|
14
|
+
Arrow::Decimal128ArrayBuilder.build(
|
15
|
+
@arrow_type,
|
16
|
+
values.map do |value|
|
17
|
+
Arrow::Decimal128.new(format(@format_str, BigDecimal(value)))
|
18
|
+
end
|
19
|
+
)
|
20
|
+
end
|
21
|
+
|
22
|
+
def arrow_type_builder
|
23
|
+
Arrow::Decimal128DataType.new(
|
24
|
+
@precision, @scale
|
25
|
+
)
|
26
|
+
end
|
27
|
+
end
|
28
|
+
end
|
29
|
+
end
|
@@ -0,0 +1,29 @@
|
|
1
|
+
# frozen_string_literal: true
|
2
|
+
|
3
|
+
module Parqueteur
|
4
|
+
module Types
|
5
|
+
class Decimal256Type < Parqueteur::Type
|
6
|
+
def initialize(options = {}, &block)
|
7
|
+
@scale = options.fetch(:scale)
|
8
|
+
@precision = options.fetch(:precision)
|
9
|
+
@format_str = "%.#{@scale}f"
|
10
|
+
super(options, &block)
|
11
|
+
end
|
12
|
+
|
13
|
+
def build_value_array(values)
|
14
|
+
Arrow::Decimal256ArrayBuilder.build(
|
15
|
+
@arrow_type,
|
16
|
+
values.map do |value|
|
17
|
+
Arrow::Decimal256.new(format(@format_str, BigDecimal(value)))
|
18
|
+
end
|
19
|
+
)
|
20
|
+
end
|
21
|
+
|
22
|
+
def arrow_type_builder
|
23
|
+
Arrow::Decimal256DataType.new(
|
24
|
+
@precision, @scale
|
25
|
+
)
|
26
|
+
end
|
27
|
+
end
|
28
|
+
end
|
29
|
+
end
|
@@ -0,0 +1,19 @@
|
|
1
|
+
# frozen_string_literal: true
|
2
|
+
|
3
|
+
module Parqueteur
|
4
|
+
module Types
|
5
|
+
class Time32Type < Parqueteur::Type
|
6
|
+
def build_value_array(values)
|
7
|
+
Arrow::Time32Array.new(
|
8
|
+
@options.fetch(:precision, :second), values
|
9
|
+
)
|
10
|
+
end
|
11
|
+
|
12
|
+
def arrow_type_builder
|
13
|
+
Arrow::Time32DataType.new(
|
14
|
+
options.fetch(:unit, :second)
|
15
|
+
)
|
16
|
+
end
|
17
|
+
end
|
18
|
+
end
|
19
|
+
end
|
@@ -0,0 +1,19 @@
|
|
1
|
+
# frozen_string_literal: true
|
2
|
+
|
3
|
+
module Parqueteur
|
4
|
+
module Types
|
5
|
+
class Time64Type < Parqueteur::Type
|
6
|
+
def build_value_array(values)
|
7
|
+
Arrow::Time64Array.new(
|
8
|
+
@options.fetch(:precision, :second), values
|
9
|
+
)
|
10
|
+
end
|
11
|
+
|
12
|
+
def arrow_type_builder
|
13
|
+
Arrow::Time64DataType.new(
|
14
|
+
options.fetch(:unit, :second)
|
15
|
+
)
|
16
|
+
end
|
17
|
+
end
|
18
|
+
end
|
19
|
+
end
|
data/lib/parqueteur/version.rb
CHANGED
data/lib/parqueteur.rb
CHANGED
@@ -2,9 +2,10 @@
|
|
2
2
|
|
3
3
|
require 'json'
|
4
4
|
require 'singleton'
|
5
|
+
require 'tempfile'
|
6
|
+
require 'parquet'
|
5
7
|
|
6
|
-
require_relative
|
7
|
-
require 'parqueteur/chunked_converter'
|
8
|
+
require_relative 'parqueteur/version'
|
8
9
|
require 'parqueteur/column'
|
9
10
|
require 'parqueteur/column_collection'
|
10
11
|
require 'parqueteur/converter'
|
@@ -14,16 +15,20 @@ require 'parqueteur/type'
|
|
14
15
|
require 'parqueteur/type_resolver'
|
15
16
|
require 'parqueteur/types/array_type'
|
16
17
|
require 'parqueteur/types/boolean_type'
|
18
|
+
require 'parqueteur/types/date32_type'
|
19
|
+
require 'parqueteur/types/date64_type'
|
20
|
+
require 'parqueteur/types/decimal128_type'
|
21
|
+
require 'parqueteur/types/decimal256_type'
|
17
22
|
require 'parqueteur/types/int32_type'
|
18
23
|
require 'parqueteur/types/int64_type'
|
19
24
|
require 'parqueteur/types/map_type'
|
20
25
|
require 'parqueteur/types/string_type'
|
21
26
|
require 'parqueteur/types/struct_type'
|
27
|
+
require 'parqueteur/types/time32_type'
|
28
|
+
require 'parqueteur/types/time64_type'
|
22
29
|
require 'parqueteur/types/timestamp_type'
|
23
|
-
require 'parquet'
|
24
30
|
|
25
31
|
module Parqueteur
|
26
32
|
class Error < StandardError; end
|
27
33
|
class TypeNotFound < Error; end
|
28
|
-
# Your code goes here...
|
29
34
|
end
|
data/parqueteur.gemspec
CHANGED
@@ -8,8 +8,8 @@ Gem::Specification.new do |spec|
|
|
8
8
|
spec.authors = ["Julien D."]
|
9
9
|
spec.email = ["julien@pocketsizesun.com"]
|
10
10
|
spec.license = 'Apache-2.0'
|
11
|
-
spec.summary = 'Parqueteur - A Ruby gem that convert
|
12
|
-
spec.description = 'Convert
|
11
|
+
spec.summary = 'Parqueteur - A Ruby gem that convert data to Parquet'
|
12
|
+
spec.description = 'Convert data to Parquet'
|
13
13
|
spec.homepage = 'https://github.com/pocketsizesun/parqueteur-ruby'
|
14
14
|
spec.required_ruby_version = Gem::Requirement.new(">= 2.3.0")
|
15
15
|
|
@@ -0,0 +1,18 @@
|
|
1
|
+
#!/bin/sh
|
2
|
+
|
3
|
+
if [ $(dpkg-query -W -f='${Status}' apache-arrow-apt-source 2>/dev/null | grep -c "ok installed") -eq 1 ]
|
4
|
+
then
|
5
|
+
exit 0
|
6
|
+
fi
|
7
|
+
|
8
|
+
LSB_RELEASE_CODENAME_SHORT=$(lsb_release --codename --short)
|
9
|
+
|
10
|
+
apt-get update
|
11
|
+
apt-get install -y -V ca-certificates lsb-release wget
|
12
|
+
wget https://apache.jfrog.io/artifactory/arrow/$(lsb_release --id --short | tr 'A-Z' 'a-z')/apache-arrow-apt-source-latest-$(lsb_release --codename --short).deb
|
13
|
+
apt-get install -y -V ./apache-arrow-apt-source-latest-${LSB_RELEASE_CODENAME_SHORT}.deb
|
14
|
+
rm ./apache-arrow-apt-source-latest-${LSB_RELEASE_CODENAME_SHORT}.deb
|
15
|
+
apt-get update
|
16
|
+
apt-get install -y libgirepository1.0-dev libarrow-dev libarrow-glib-dev libparquet-dev libparquet-glib-dev
|
17
|
+
|
18
|
+
exit 0
|
metadata
CHANGED
@@ -1,14 +1,14 @@
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
2
2
|
name: parqueteur
|
3
3
|
version: !ruby/object:Gem::Version
|
4
|
-
version: 1.1
|
4
|
+
version: 1.3.1
|
5
5
|
platform: ruby
|
6
6
|
authors:
|
7
7
|
- Julien D.
|
8
8
|
autorequire:
|
9
9
|
bindir: exe
|
10
10
|
cert_chain: []
|
11
|
-
date: 2021-10-
|
11
|
+
date: 2021-10-17 00:00:00.000000000 Z
|
12
12
|
dependencies:
|
13
13
|
- !ruby/object:Gem::Dependency
|
14
14
|
name: red-parquet
|
@@ -24,7 +24,7 @@ dependencies:
|
|
24
24
|
- - "~>"
|
25
25
|
- !ruby/object:Gem::Version
|
26
26
|
version: '5.0'
|
27
|
-
description: Convert
|
27
|
+
description: Convert data to Parquet
|
28
28
|
email:
|
29
29
|
- julien@pocketsizesun.com
|
30
30
|
executables: []
|
@@ -38,9 +38,15 @@ files:
|
|
38
38
|
- Rakefile
|
39
39
|
- bin/console
|
40
40
|
- bin/setup
|
41
|
-
-
|
41
|
+
- examples/cars.rb
|
42
|
+
- examples/convert-and-compression.rb
|
43
|
+
- examples/convert-methods.rb
|
44
|
+
- examples/convert-to-io.rb
|
45
|
+
- examples/convert-with-chunks.rb
|
46
|
+
- examples/convert-without-compression.rb
|
47
|
+
- examples/hello-world.rb
|
48
|
+
- examples/readme-example.rb
|
42
49
|
- lib/parqueteur.rb
|
43
|
-
- lib/parqueteur/chunked_converter.rb
|
44
50
|
- lib/parqueteur/column.rb
|
45
51
|
- lib/parqueteur/column_collection.rb
|
46
52
|
- lib/parqueteur/converter.rb
|
@@ -50,15 +56,21 @@ files:
|
|
50
56
|
- lib/parqueteur/type_resolver.rb
|
51
57
|
- lib/parqueteur/types/array_type.rb
|
52
58
|
- lib/parqueteur/types/boolean_type.rb
|
59
|
+
- lib/parqueteur/types/date32_type.rb
|
60
|
+
- lib/parqueteur/types/date64_type.rb
|
61
|
+
- lib/parqueteur/types/decimal128_type.rb
|
62
|
+
- lib/parqueteur/types/decimal256_type.rb
|
53
63
|
- lib/parqueteur/types/int32_type.rb
|
54
64
|
- lib/parqueteur/types/int64_type.rb
|
55
65
|
- lib/parqueteur/types/map_type.rb
|
56
66
|
- lib/parqueteur/types/string_type.rb
|
57
67
|
- lib/parqueteur/types/struct_type.rb
|
68
|
+
- lib/parqueteur/types/time32_type.rb
|
69
|
+
- lib/parqueteur/types/time64_type.rb
|
58
70
|
- lib/parqueteur/types/timestamp_type.rb
|
59
71
|
- lib/parqueteur/version.rb
|
60
72
|
- parqueteur.gemspec
|
61
|
-
-
|
73
|
+
- scripts/apache-arrow-ubuntu-install.sh
|
62
74
|
homepage: https://github.com/pocketsizesun/parqueteur-ruby
|
63
75
|
licenses:
|
64
76
|
- Apache-2.0
|
@@ -82,5 +94,5 @@ requirements: []
|
|
82
94
|
rubygems_version: 3.2.3
|
83
95
|
signing_key:
|
84
96
|
specification_version: 4
|
85
|
-
summary: Parqueteur - A Ruby gem that convert
|
97
|
+
summary: Parqueteur - A Ruby gem that convert data to Parquet
|
86
98
|
test_files: []
|
data/example.rb
DELETED
@@ -1,39 +0,0 @@
|
|
1
|
-
require 'bundler/setup'
|
2
|
-
require 'parqueteur'
|
3
|
-
|
4
|
-
class Foo < Parqueteur::Converter
|
5
|
-
column :id, :bigint
|
6
|
-
column :reference, :string
|
7
|
-
column :hash, :map, key: :string, value: :string
|
8
|
-
column :valid, :boolean
|
9
|
-
column :total, :integer
|
10
|
-
column :numbers, :array, elements: :integer
|
11
|
-
column :my_struct, :struct do
|
12
|
-
field :test, :string
|
13
|
-
field :mon_nombre, :integer
|
14
|
-
end
|
15
|
-
end
|
16
|
-
|
17
|
-
LETTERS = ('a'..'z').to_a
|
18
|
-
|
19
|
-
data = 1000.times.collect do |i|
|
20
|
-
{
|
21
|
-
'id' => i + 1,
|
22
|
-
'reference' => "coucou:#{i}",
|
23
|
-
'hash' => { 'a' => LETTERS.sample },
|
24
|
-
'valid' => rand < 0.5,
|
25
|
-
'total' => rand(100..500),
|
26
|
-
'numbers' => [1, 2, 3],
|
27
|
-
'my_struct' => {
|
28
|
-
'test' => 'super'
|
29
|
-
}
|
30
|
-
}
|
31
|
-
end
|
32
|
-
|
33
|
-
# chunked_converter = Parqueteur::ChunkedConverter.new(data, Foo)
|
34
|
-
# pp chunked_converter.write_files('test')
|
35
|
-
puts Foo.convert(data, output: 'tmp/test.parquet')
|
36
|
-
table = Arrow::Table.load('tmp/test.parquet')
|
37
|
-
table.each_record do |record|
|
38
|
-
puts record.to_h
|
39
|
-
end
|
@@ -1,28 +0,0 @@
|
|
1
|
-
# frozen_string_literal: true
|
2
|
-
|
3
|
-
module Parqueteur
|
4
|
-
class ChunkedConverter
|
5
|
-
attr_reader :schema
|
6
|
-
|
7
|
-
def initialize(input, converter, chunk_size = 200)
|
8
|
-
@input = Parqueteur::Input.from(input)
|
9
|
-
@converter = converter
|
10
|
-
@chunk_size = chunk_size
|
11
|
-
end
|
12
|
-
|
13
|
-
def chunks
|
14
|
-
Enumerator.new do |arr|
|
15
|
-
@input.each_slice(@chunk_size) do |chunk|
|
16
|
-
local_converter = @converter.new(chunk)
|
17
|
-
arr << local_converter.to_io
|
18
|
-
end
|
19
|
-
end
|
20
|
-
end
|
21
|
-
|
22
|
-
def write_files(prefix)
|
23
|
-
chunks.each_with_index do |chunk, idx|
|
24
|
-
File.write("#{prefix}.#{idx}.parquet", chunk.read)
|
25
|
-
end
|
26
|
-
end
|
27
|
-
end
|
28
|
-
end
|
data/test.json
DELETED
@@ -1 +0,0 @@
|
|
1
|
-
[{"id":1,"reference":"coucou","hash":{"a":"b"},"valid":true,"hash2":{},"numbers":[1,2,3],"map_array":[]},{"id":2,"reference":"coucou","hash":{"c":"d"},"valid":false,"hash2":{},"numbers":[4,5,6],"map_array":[]},{"id":3,"reference":"coucou","hash":{"e":"f"},"valid":true,"hash2":{"x":[1,2,3]},"numbers":[7,8,9],"map_array":[{"x":"y"}]}]
|