Skip to content

A data generator source connector for Flink SQL based on data-faker.

License

Notifications You must be signed in to change notification settings

knaufk/flink-faker

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Build Status

flink-faker

flink-faker is an Apache Flink table source that generates fake data based on the Data Faker expression provided for each column.

Checkout this demo web application for some example Java Faker (fully compatible with Data Faker) expressions and Data Faker documentation.

This project is inspired by voluble.

Package

mvn clean package

Compatibility Matrix

Flink Version flink-faker Version
1.11 0.1.x - 0.4.x
1.12 0.1.x - 0.4.x
1.13 0.1.x - 0.4.x
1.14 0.1.x - 0.4.x
1.15 0.5.0
1.16 0.5.1+
1.17 0.5.1+

There are no automated tests that check this compatibility. So, please treat this table as "best knowledge". If you notice any incompatibilities please open an issue.

Using flink-faker with the Flink SQL Client

  1. Download Flink from the Apache Flink website.
  2. Download the flink-faker JAR from the Releases page (or build it yourself).
  3. Put the downloaded jars under lib/.
  4. (Re)Start a Flink cluster.
  5. (Re)Start the Flink CLI.

Usage

As ScanTableSource

CREATE TEMPORARY TABLE heros (
  `name` STRING,
  `power` STRING, 
  `age` INT
) WITH (
  'connector' = 'faker', 
  'fields.name.expression' = '#{superhero.name}',
  'fields.power.expression' = '#{superhero.power}',
  'fields.power.null-rate' = '0.05',
  'fields.age.expression' = '#{number.numberBetween ''0'',''1000''}'
);
SELECT * FROM heros;

As LookupTableSource

CREATE TEMPORARY TABLE location_updates (
  `character_id` INT,
  `location` STRING,
  `proctime` AS PROCTIME()
)
WITH (
  'connector' = 'faker', 
  'fields.character_id.expression' = '#{number.numberBetween ''0'',''100''}',
  'fields.location.expression' = '#{harry_potter.location}'
);
CREATE TEMPORARY TABLE characters (
  `character_id` INT,
  `name` STRING
)
WITH (
  'connector' = 'faker', 
  'fields.character_id.expression' = '#{number.numberBetween ''0'',''100''}',
  'fields.name.expression' = '#{harry_potter.characters}'
);
SELECT 
  c.character_id,
  l.location,
  c.name
FROM location_updates AS l
JOIN characters FOR SYSTEM_TIME AS OF proctime AS c
ON l.character_id = c.character_id;

Currently, the faker source supports the following data types:

  • CHAR
  • VARCHAR
  • STRING
  • TINYINT
  • SMALLINT
  • INTEGER
  • BIGINT
  • FLOAT
  • DOUBLE
  • DECIMAL
  • BOOLEAN
  • TIMESTAMP
  • DATE
  • TIME
  • ARRAY
  • MAP
  • MULTISET
  • ROW

Connector Options

Connector Option Default Description
number-of-rows None The number of rows to produce. If this is options is set, the source is bounded otherwise it is unbounded and runs indefinitely.
rows-per-second 10000 The maximum rate at which the source produces records.
fields.<field>.expression None The Data Faker expression to generate the values for this field.
fields.<field>.null-rate 0.0 Fraction of rows for which this field is null
fields.<field>.length 1 Size of array, map or multiset

On Timestamps

For rows of type TIMESTAMP, DATE the corresponding Data Faker expression needs to return a timestamp formatted as uuuu-MM-dd hh:mi:ss[.nnnnnnnnn]. Typically, you would use one of the following expressions:

CREATE TEMPORARY TABLE timestamp_time_and_date_example (
  `timestamp1` TIMESTAMP(3),
  `timestamp2` TIMESTAMP(3),
  `timestamp3` TIMESTAMP(3),
  `time`       TIME,
  `date1`      DATE,
  `date2`      DATE
)
WITH (
  'connector' = 'faker', 
  'fields.timestamp1.expression' = '#{date.past ''15'',''SECONDS''}',
  'fields.timestamp2.expression' = '#{date.past ''15'',''5'',''SECONDS''}',
  'fields.timestamp3.expression' = '#{date.future ''15'',''5'',''SECONDS''}',
  'fields.time.expression' = '#{time.future ''15'',''5'',''SECONDS''}',
  'fields.date1.expression' = '#{date.birthday}',
  'fields.date2.expression' = '#{date.birthday ''1'',''100''}'
);
SELECT * FROM timestamp_time_and_date_example;

For timestamp1 Data Faker will generate a random timestamp that lies at most 15 seconds in the past. For timestamp2 Data Faker will generate a random timestamp, that lies at most 15 seconds in the past, but at least 5 seconds. For timestamp3 Data Faker will generate a random timestamp, that lies at most 15 seconds in the future, but at least 5 seconds. For time Data Faker will generate a random time, that lies at most 15 seconds in the future, but at least 5 seconds. For date1 Data Faker will generate a random birthday between 18 and 65 years ago. For date2 Data Faker will generate a random birthday between 1 and 100 years ago.

On Collection Data Types

The usage of ARRAY, MULTISET, MAP and ROW types is shown in the following example.

CREATE TEMPORARY TABLE hp (
  `character-with-age` MAP<STRING,INT>,
  `spells` MULTISET<STRING>,
  `locations` ARRAY<STRING>,
  `house-points` ROW<`house` STRING, `points` INT>
) WITH (
  'connector' = 'faker',
  'fields.character-with-age.key.expression' = '#{harry_potter.character}',
  'fields.character-with-age.value.expression' = '#{number.numberBetween ''10'',''100''}',
  'fields.character-with-age.length' = '2',
  'fields.spells.expression' = '#{harry_potter.spell}',
  'fields.spells.length' = '5',
  'fields.locations.expression' = '#{harry_potter.location}',
  'fields.locations.length' = '3',
  'fields.house-points.house.expression' = '#{harry_potter.house}',
  'fields.house-points.points.expression' = '#{number.numberBetween ''10'',''100''}'
);
SELECT * FROM hp;

"One Of" Columns

Datafaker allows to pick a random value from a list of options via expression Options.option

CREATE TEMPORARY TABLE orders (
  `order_id` INT,
  `order_status` STRING
)
WITH (
  'connector' = 'faker',
  'fields.order_id.expression' = '#{number.numberBetween ''0'',''100''}',
  'fields.order_status.expression' = '#{Options.option ''RECEIVED'',''SHIPPED'',''CANCELLED'')}'
);
SELECT * FROM orders;

License

Copyright © 2020-2023 Konstantin Knauf

Distributed under Apache License, Version 2.0.