Any optimization that I can do to improve execute_field performance?

[junyuanz1]

our schema configuration looks like this:

1. we use lazy loader to load an object
2. and we need to return all the fields from that object

we noticed that the ruby-graphql has to call execute_field on every field one by one
since our field is pretty big, it ends up taking a long time to go thought all the fields

the graphql gem that i am using is very old 1.13.18 not sure whether newer version has any optimization around this or not.

[rmosolgo]

Hi! There's an internal wrapper called RawValue which GraphQL-Ruby inserts without modification into the response whenever it's found:

graphql-ruby/lib/graphql/execution/interpreter/runtime.rb

Lines 546 to 549 in 24f797a

	when GraphQL::Execution::Interpreter::RawValue
	# Write raw value directly to the response without resolving nested objects
	set_result(selection_result, result_name, value.resolve, false, is_non_null)
	HALT

You could try using that if you don't need any more GraphQL resolution. (But you have to handle _subfields, too, if there are any.) That's all I can think of!

To dig in further, please open an issue with some profiling data, for example: https://graphql-ruby.org/testing/profiling.html