vinibs opened a new issue, #21039:
URL: https://github.com/apache/superset/issues/21039

   ## [SIP-88] Proposal for a dynamic number of bins in Histogram
   
   ### Motivation
   
   At my company, we started to use Superset to generate some charts for our 
data and we faced troubles when using the Histogram visualization due to its 
number of bins: our data, for the same time interval, could bring values whose 
ranges can vary a lot (in a scenario, they could be from 1 to 3, in another, 
from 1 to 27, for example) and we need the bins to be exactly one for each 
value - if our interval is from 1 to 5, we need 5 bins; if it's from 1 to 27, 
we need 27 bins. We can only set a fixed number of bins regardless of which is 
our data interval.
   
   ### Proposed Change
   
   We thought about two possible ways to solve it: the first one would be 
setting, as the Histogram's default number of bins, the `MAX()` value in the 
data range, so, if we haven't defined a fixed number of bins, it would make a 
histogram with one bin for each value of our data; The second approach would be 
to allow to set a number of bins based on a query (for example, this `MAX()` 
function over the generated data) to dynamically set the number of bins based 
on our data's range.
   For the second approach, it would involve changing the "Number of bins" 
select input, probably by also adding an extra window to insert the query we 
need to use on it.
   
   ### New or Changed Public Interfaces
   
   This would affect mostly the Histogram visualization and the "Number of 
bins" selection component, in terms of User Interface and mainly the behavior 
of the bin calculation when generating the visualization.
   
   ### New dependencies
   
   We don't think it would be necessary to add any new dependency.
   
   ### Migration Plan and Compatibility
   
   At first, we don't see a need for any migration.
   
   ### Rejected Alternatives
   
   We didn't think of any alternative that was already been rejected. The idea 
is to bring some discussion to find the best approach here.
   
   We're planning to try to implement it here to solve our problem as quickly 
as possible and, if this solution is good enough to merge here as an 
improvement, we'd be happy to do so.
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]


---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to