BigQuery: expose public helper method to convert a list of schema fields to/from a list of schema dictionaries (JSON)

[tswast]

I've received some feedback asking how to take a schema from the client library and save it to a JSON file. One reason to do this is the bq command-line tool accepts a schema file, formatted like

[
  {"name": "name", "type": "string", "mode": "required"},
  {"name": "gender", "type": "string", "mode": "nullable"},
  {"name": "count", "type": "integer", "mode": "required"}
]

Note: this format is the same as the API representation.

It would be great if our client libraries could read/write in this format.

[tseaver]

We already have non-public helper methods which render / parse schemas to / from a format like:

{"fields": [
  {"name": "name", "type": "string", "mode": "required"},
  {"name": "gender", "type": "string", "mode": "nullable"},
  {"name": "count", "type": "integer", "mode": "required"}
]}

It would be pretty trivial to adapt them to skip / overlook the missing fields key. How would you envision this for the library API surface? Methods on the Table class to load / store the schema to a file? What about schema for e.g. query jobs?

[tswast]

There are two main use cases that I've heard about

• at table creation/load time, being able to load schema from a file.
• after a query so that they can serialize the schema from the stats to then use to create other table.

[tswast]

Since this is something that would act on lists, it doesn't really correspond to an existing resource representation that we have. Do we put static helper functions in the top-level google.cloud.apiname package, generally?

[tswast]

Going through the issues I see that this is a duplicate of #3419.

Closing this issue in favor of that one.