GetSchema

Warnings

• breaking Versions prior to 0.2.5 incorrectly auto-converted values like 0, 0.0, empty strings (''), and false to `null` during schema inference. This behavior was a bug and has been fixed in v0.2.5 and later.
  Fix: Upgrade to getschema v0.2.5 or later. If your logic previously relied on this implicit conversion, ensure your input data is pre-processed or adjust your schema expectations accordingly.
• gotcha The default inference for fields containing only `null` values changed in v0.2.10. Prior to this version, such fields might have been inferred as `{"type": "null"}` or `{"type": "string"}` (since v0.2.3). From v0.2.10, they default to `{"type": ["null", "string"]}` (nullable string).
  Fix: If 'nullable string' is not the desired type for fields containing only nulls, explicitly define the expected type in a post-processing step or through custom logic.
• gotcha In versions prior to v0.2.3, if a field contained only `null` values across all sample records, the inferred JSON Schema type for that field would be `{"type": "null"}`. From v0.2.3, this default was changed to `{"type": "string"}` (and later to `{"type": ["null", "string"]}` in v0.2.10).
  Fix: Review schemas inferred from fields with only null values if using older versions. Upgrade to a newer version for consistent handling, or explicitly specify the desired type for such fields.

Install

• pip install getschema Install with pip

Imports

• infer_schema

  import getschema
  # or
  from getschema import infer_schema

Quickstart

This quickstart demonstrates how to use `getschema.infer_schema` to generate a JSON Schema from a list of Python dictionaries representing sample records. The output is a formatted JSON string of the inferred schema.

import getschema
import json

sample_records = [
    {"name": "Alice", "age": 30, "city": "New York"},
    {"name": "Bob", "age": 24, "city": "London", "email": "bob@example.com"},
    {"name": "Charlie", "age": None, "city": "Paris"},
    {"name": "David", "age": 35, "city": "Berlin", "hobbies": ["reading", "hiking"]}
]

# Infer schema from a list of records
schema = getschema.infer_schema(sample_records)
print(json.dumps(schema, indent=2))