holidayextras/bigjson

BigJson

A small library to parse JSON schemas and output to big query schema only, for now.

Installation

pip3 install -r requirements.txt

Usage

Consuming

Convert schema

from bigjson import convert

print(convert(json))

CLI

Check schema

Do not specify a GCloud project.

cat schema.json | python3 bigjson/__init__.py

Create table

Specify a GCloud project.

cat schema.json | python3 bigjson/__init__.py project-id dataset-id

bigjson's Issues

big query cannot take multiple types like jsonschema can

In jsonschema we can easily represent this and it be valid

      "description": "would prefer value here for consistency with other tracking",
      "type": [
        "object",
        "string"
      ]
    }

However in big query you can only have one type. It would be easy to force end user to stringify their data and post process it. Would you need to set this explicitly or implicitly? I feel like the mroe info the better

jsonchema using only pattern

jsonschema allows a schema to have only the pattern to recognise that it therefore must be string. There currently is no login in this repo to assume that if type is not present but we have pattern he can assume its string.

Should be a quick fix