Function arrow::json::reader::infer_json_schema_from_iterator

pub fn infer_json_schema_from_iterator<I>(value_iter: I) -> Result<Schema> where
    I: Iterator<Item = Result<Value>>, 

Infer the fields of a JSON file by reading all items from the JSON Value Iterator.

The following type coercion logic is implemented:

• Int64 and Float64 are converted to Float64
• Lists and scalars are coerced to a list of a compatible scalar
• All other cases are coerced to Utf8 (String)

Note that the above coercion logic is different from what Spark has, where it would default to String type in case of List and Scalar values appeared in the same field.

The reason we diverge here is because we don’t have utilities to deal with JSON data once it’s interpreted as Strings. We should match Spark’s behavior once we added more JSON parsing kernels in the future.