secondary axis support

[jerry0715]

Hi,

I just try the Brunel from jupyter. It's amazing.

I use a sample CSV to draw a line chart with a secondary aix which the value is much smaller than the primary value.

Does Brunel already support it or in later release?
Thanks in advance.

[grahamwills]

Currently this is not supported; we had a version which "mostly" worked in prototype, but dropped the support as it was too buggy. One question we had was how we handled the decision to use a single merged axis or to create two axes.

Option A: Handle automatically -- maybe a rule like "if each unmarked axis is at least 25% of the size of the merged, use the merged, otherwise use two axes"
Option B: Modify the '+' operator -- maybe "++" means second axis (but then you could not control X and Y separately)
Option C: Add an option to the specification of y (and x) ... maybe something like "y(avg_rt:dual)"

At the moment I'm leaning towards A and C -- Implement a useful rule, but allow user override through the coordinate definitions. Please add any thoughts you might have ...

[jerry0715]

Thanks a lot for your reply.
Personally Option C is a good idea for me.

BTW, I have an immature idea that if the data/sample of X axis could be mapped into same range, is there a way to put different Y axis layered on each other?

For example, in this case, iops and avg_rt(average responsive time) share same time points and have different Y. The layer of The X(time) - Y(avg_rt) could be added on layer of X(time) -Y(iops). Maybe it is also possible to add a third X(time) - Y(bandwidth) layer to above two layers.

[micheusch]

option C would be ace, thanks!