JSONs to Looker views (J2V)
J2V is a simple command-line tool to convert JSON to Looker readable files in forms of Looker Views and Looker Explores.
Also it outputs an SQL with proper paths and explosion expressions.
This is useful to be used in combination with databases that are focusing on schema-on-read, and data is stored in raw JSON instead of exploded into columns of a table or view.
Example use case
You have a table in your database. This table contains a column containing JSONs (one JSON per row). You are very curious how these data look like exploded, but you do not want to spend 2h going through the JSON structure and specifying all the fields just to surface them in Looker.
With J2V all the structures are discovered automatically and two files are generated - a Looker View and Looker Explore. All you need to do is copy/paste the output of this command line tool into your Looker project and you can start exploring.
Usage
Requirements
Python 3 must be installed.
How to run
- use code from github or
pip install j2v
Parameters
-
json_files: Files in JSON format, representing the data stored in a table -
output_view: Name of Looker View output file to be created -
output_explore: Name of Looker model output file to be created -
sql_table_name: Name of the DB table to be used (this is only used in the LookML files; no actual connection to a database will be done as part of this tool) -
table_alias: Name of the table alias -
column_name: Name of the column in the DB table as specified insql_table_name. (this is only used in the LookML files; no actual connection to a database will be done as part of this tool) -
primary_key: Name of the primary key from JSON field
Output
-
output_view: File containing definitions of Looker views (see examples directory in this repository) -
output_explore: File containing definition of looker explore exploding the structures (see examples directory in this repository)
Example usage
Using all parameters
python main.py --json_files data1.json data2.json --output_view RESTAURANT_CHAIN --output_explore RESTAURANT_CHAIN --column_name DATA --sql_table_name RESTAURANT_DETAILS --table_alias chains_table --handle_null_values_in_sql true --primary_key apiVersion
Using only mandatory parameters
python3 main.py --json_files order_example.json order_example2.json order_example3.json
Contribution
Project structure:
-
j2v- source code of a package -
examples- working examples -
tests- tests
Contribute
- If unsure, open an issue for a discussion
- Create a fork
- Make your change
- Make a pull request
- Happy contribution!
EXAMPLE
Input:
{
"apiVersion": "v3.4",
"data Provider": "Eat me",
"restaurants": [
{
"name": "Super Burger",
"city": "Sydney",
"country": "Australia",
"address": "Big Street 3",
"currency": "AUD",
"openTime": 1571143824,
"menu": [
{
"dishName": "BurgerPlus",
"price": 10,
"ingredients": ["Meat", "Cheese", "Bun"]
}
]
}
],
"headquarter": {
"employees": 36,
"city": "Olsztyn",
"country": "Poland",
"building": {
"address": "3 Maja 10",
"floors": [1, 2, 7]
}
},
"dataGenerationTimestamp": "2019-03-30T11:30:00.812Z",
"payloadPrimaryKeyValue": "3ab21b54-22d6-473c-b055-4430f8927d4c",
"version": null
}Ouput:
SQL output (now only Snowflake dialect supported):
---VIEW WITH NUll VALUE HANDLING---
SELECT
---chains_table Information
IFNULL(chains_table."DATA":"apiVersion"::string,'N/A') AS API_VERSION,
IFNULL(chains_table."DATA":"data Provider"::string,'N/A') AS DATA_PROVIDER,
IFNULL(chains_table."DATA":"headquarter":"building":"address"::string,'N/A') AS HEADQUARTER_BUILDING_ADDRESS,
IFNULL(chains_table."DATA":"headquarter":"city"::string,'N/A') AS HEADQUARTER_CITY,
IFNULL(chains_table."DATA":"headquarter":"country"::string,'N/A') AS HEADQUARTER_COUNTRY,
IFNULL(chains_table."DATA":"headquarter":"employees"::number,0) AS HEADQUARTER_EMPLOYEES,
IFNULL(chains_table."DATA":"payloadPrimaryKeyValue"::string,'N/A') AS PAYLOAD_PRIMARY_KEY_VALUE,
IFNULL(chains_table."DATA":"version"::string,'N/A') AS VERSION,
chains_table."DATA":"dataGenerationTimestamp"::timestamp AS DATA_GENERATION_TIMESTAMP,
---restaurants Information
IFNULL(restaurants.VALUE:"address"::string,'N/A') AS RESTAURANTS_ADDRESS,
IFNULL(restaurants.VALUE:"city"::string,'N/A') AS RESTAURANTS_CITY,
IFNULL(restaurants.VALUE:"country"::string,'N/A') AS RESTAURANTS_COUNTRY,
IFNULL(restaurants.VALUE:"currency"::string,'N/A') AS RESTAURANTS_CURRENCY,
IFNULL(restaurants.VALUE:"name"::string,'N/A') AS RESTAURANTS_NAME,
IFNULL(restaurants.VALUE:"openTime"::number,0) AS RESTAURANTS_OPEN_TIME,
---restaurants_menu Information
IFNULL(restaurants_menu.VALUE:"dishName"::string,'N/A') AS RESTAURANTS_MENU_DISH_NAME,
IFNULL(restaurants_menu.VALUE:"price"::number,0) AS RESTAURANTS_MENU_PRICE,
---restaurants_menu_ingredients Information
IFNULL(restaurants_menu_ingredients.VALUE::string,'N/A') AS RESTAURANTS_MENU_INGREDIENTS_VALUE,
---headquarter_building_floors Information
IFNULL(headquarter_building_floors.VALUE::number,0) AS HEADQUARTER_BUILDING_FLOORS_VALUE
FROM RESTAURANT_DETAILS AS chains_table,
LATERAL FLATTEN(OUTER => TRUE, INPUT => chains_table."DATA":"restaurants") restaurants,
LATERAL FLATTEN(OUTER => TRUE, INPUT => restaurants.VALUE:"menu") restaurants_menu,
LATERAL FLATTEN(OUTER => TRUE, INPUT => restaurants_menu.VALUE:"ingredients") restaurants_menu_ingredients,
LATERAL FLATTEN(OUTER => TRUE, INPUT => chains_table."DATA":"headquarter":"building":"floors") headquarter_building_floorsOuput files:
View file:
view: chains_table {
sql_table_name: RESTAURANT_DETAILS ;;
dimension: address {
description: "Address"
type: string
sql: ${TABLE}."DATA":"headquarter":"building":"address"::string ;;
group_label: "Building"
}
dimension: api_version {
description: "Api version"
primary_key: yes
type: string
sql: ${TABLE}."DATA":"apiVersion"::string ;;
}
dimension: city {
description: "City"
type: string
sql: ${TABLE}."DATA":"headquarter":"city"::string ;;
group_label: "Headquarter"
}
dimension: country {
description: "Country"
type: string
sql: ${TABLE}."DATA":"headquarter":"country"::string ;;
group_label: "Headquarter"
}
dimension: data_provider {
description: "Data provider"
type: string
sql: ${TABLE}."DATA":"data Provider"::string ;;
}
dimension: employees {
description: "Employees"
type: number
sql: ${TABLE}."DATA":"headquarter":"employees"::number ;;
group_label: "Headquarter"
}
dimension: payload_primary_key_value {
description: "Payload primary key value"
type: string
sql: ${TABLE}."DATA":"payloadPrimaryKeyValue"::string ;;
}
dimension: version {
description: "Version"
type: string
sql: ${TABLE}."DATA":"version"::string ;;
}
dimension_group: data_generation_timestamp {
description: "Data generation timestamp"
type: time
timeframes: [
raw,
time,
date,
week,
month,
quarter,
year
]
sql: ${TABLE}."DATA":"dataGenerationTimestamp"::timestamp ;;
}
}
view: restaurants {
dimension: address {
description: "Address"
type: string
sql: ${TABLE}.VALUE:"address"::string ;;
}
dimension: city {
description: "City"
type: string
sql: ${TABLE}.VALUE:"city"::string ;;
}
dimension: country {
description: "Country"
type: string
sql: ${TABLE}.VALUE:"country"::string ;;
}
dimension: currency {
description: "Currency"
type: string
sql: ${TABLE}.VALUE:"currency"::string ;;
}
dimension: name {
description: "Name"
type: string
sql: ${TABLE}.VALUE:"name"::string ;;
}
dimension_group: open_time {
description: "Open time"
datatype: epoch
type: time
timeframes: [
raw,
time,
date,
week,
month,
quarter,
year
]
sql: ${TABLE}.VALUE:"openTime"::number ;;
}
}
view: restaurants_menu {
dimension: dish_name {
description: "Dish name"
type: string
sql: ${TABLE}.VALUE:"dishName"::string ;;
}
dimension: price {
description: "Price"
type: number
sql: ${TABLE}.VALUE:"price"::number ;;
}
}
view: restaurants_menu_ingredients {
dimension: value {
description: "Value"
type: string
sql: ${TABLE}.VALUE::string ;;
}
}
view: headquarter_building_floors {
dimension: value {
description: "Value"
type: number
sql: ${TABLE}.VALUE::number ;;
}
}
Explore file:
include: "restaurant_chain.view.lkml"
explore: chains_table {
view_name: chains_table
from: chains_table
label: "chains_table explore"
description: "chains_table explore"
join: restaurants {
from: restaurants
sql:,LATERAL FLATTEN(OUTER => TRUE, INPUT => chains_table."DATA":"restaurants") restaurants;;
relationship: one_to_many
}
join: restaurants_menu {
from: restaurants_menu
sql:,LATERAL FLATTEN(OUTER => TRUE, INPUT => restaurants.VALUE:"menu") restaurants_menu;;
relationship: one_to_many
required_joins: [restaurants]
}
join: restaurants_menu_ingredients {
from: restaurants_menu_ingredients
sql:,LATERAL FLATTEN(OUTER => TRUE, INPUT => restaurants_menu.VALUE:"ingredients") restaurants_menu_ingredients;;
relationship: one_to_many
required_joins: [restaurants_menu]
}
join: headquarter_building_floors {
from: headquarter_building_floors
sql:,LATERAL FLATTEN(OUTER => TRUE, INPUT => chains_table."DATA":"headquarter":"building":"floors") headquarter_building_floors;;
relationship: one_to_many
}
}
