Custon aggregation types - or idea for enhancement

Hi,

when reading the documentation on aggregates, the current min, max, sum and avg seems,from an analytical perspective, rather limited.

  • Is there a way to custom-extend the available aggregation-functions? eg. by scripting?
  • I would currently miss median or more generally percentiles.

I know that this might have an intricate relationship on indices.

Do you have an example use case your are trying to solve? That would probably be the best way to approach this and help show exactly how it is “rather limited”. What kind of scripting do you want?

Sure. I import Log Data, this log data contains the time it takes for actions to complete in ms. I would be interested in the minimum, 25 quantil, median, 75 quantil and maximum runtime. To get a feeling of the distribution.

I could use Elastic for that, or I could export all the relevant data into R and analyze it there. But to have more aggregate functions natively supported in Dgraph would come handy.