Rick Moynihan
Rick Moynihan
We need to define an ontology which lets users annotate their data with extra hints to guide our schema generation. This should include featuers to let people specify graphql schema...
Dimensions and other types aren't shared leading to a proliferation of schema types, e.g. for gender:  Whilst for gender it's not a huge...
This has been moved from #3. e.g. Standard deviation of the measure values.
The convention we have of prefixing `dataset_` to dataset names is just a way of avoiding collisions on field names in graphql. An alternative (perhaps more elegant) approach would be...
- get observations... - [X] filtered by a set of dimension/value pairs - [ ] filtered a range of values for each the dimensions (e.g. ref period: 2001 or 2002,...
On [this pr](https://github.com/Swirrl/graphql-qb/pull/38#issuecomment-337522089) I noticed that there's a small limitation to the way we handle boolean arguments, that means they don't compose as is usually expected. e.g. Though you can...
From @zeginis over email: ``` I had a look at the code at github. However I am not very familiar to clojure. What data does the dynamically created graphql schema...
Currently the schemas are only built once at startup. We should rebuild the schemas dynamically when datasets and/or their supporting vocabularies are changed; this way we can keep in sync...
We should perhaps adopt a different convention so that non `dataset` types aren't named `dataset_` but instead had a different prefix perhaps something like `measure_type_dataset_earnings`, etc. By changing the convention...