Details
-
Improvement
-
Status: Closed
-
Major
-
Resolution: Resolved
-
None
-
None
Description
Currently the timeseries function only supports aggregations across the time dimension. This ticket will add the split parameter which will add a top level split by categorical field, to produce time lines per each split. The limit parameter will also be added and be applied only to split. The sort will always be desc on the first aggregation in the first iteration to keep the parameters simple.
Sample syntax:
timeseries(collection1, q="*:*", field="timefield", split="company", limit="10", gap="+1DAY", format="YYYY-dd-MM" , avg(price_f))
The output of this can be easily pivoted into a matrix and correlated or clustered like the output of the facet2D function. The diff and minMaxScale functions already support operations over matrix rows so it's very easy to perform clustering etc.. on this output.