-
Notifications
You must be signed in to change notification settings - Fork 417
Solution to #60 #80
base: master
Are you sure you want to change the base?
Solution to #60 #80
Conversation
Related #60. I would rather this be implemented as a wildcard operator, rather than adding a new query parameter. |
Yeah that would be nicer, I'll try hacking it out today. My idea is to only allow "type()" rather than also allow "type(prop.)" since that can be queried by "type(prop)" |
Yes, I agree: |
Sorry about the huge size, that's from me regenerating the grammar with pegjs. I had to reorder the grammar rules, looks like pegjs eagerly fails as soon as it starts matching a subrule in the grammar. |
Why not just have a query like "event" return all data for all events called "event"? |
@edmund-huber there is definitely value in the knowledge that an event simply occurred (and overhead in sending down/calculating all those properties) |
Right, we want the event expression "event" to only return the event times—it's an existence check. |
@mbostock how does this diff look? seems to work for me.. |
…k, stat.net. Events are stored with main type (up to first dot), and then a data field containing subtype. square#80
Add a parameter to /event/get to toggle retrieving all event data.
The parameter is "allData" and when present, all fields of the event are returned regardless of what is specified. (i.e. the query "expression=server(cpu)&allData" will return all fields)
While this creates another parameter rather than allowing a query like "server(*)", it is explicit in that it is a switch to return all data.