Zoom only x or y axis with mouse wheel when using Chart.interactive()?

[trllrgn]

Is there a way to detect whether the mouse is over the x axis or y axis and zoom (i.e. change scales) only on this axis with Chart.interactive()?

I found the documentation and understand that I can use bind_x=False or bind_y=False in order to stop zooming in x or y, respectively. The documentation also mentions that:

name:string The selection name to use for the axes scales. This name should be unique among all selections within the chart.

which may refer to what I need, but I could not find an example. Is what I'm looking for possible? Thanks.

[jakevdp]

No, there's currently no way to do that in Vega-Lite or Altair. The relevant feature request is here: https://github.com/vega/vega-lite/issues/1657

[trllrgn]

Thanks for the info.

[ejain]

To prevent zooming e.g. along the y axis you can specify:

.interactive(
  bind_y = False
)

The linked Vega-Lite issue seems unrelated, something about legends?

[jakevdp]

The reason the linked issue is relevant is because that's what discusses the feature in question: interactive scales and legends. Note that the original question is not about disabling y zooming everywhere, which can be done with bind_y, but rather disabling it when your mouse is hovering over the x-axis.

[LombardoAndrea195]

Is it possible to do the following issue? I would like to have to possibility to select an interval but if I do this I don't want that the x axis change. Instead when I made the zoom I would like to do that. I 've thought to do that through a combobox and when for example interval combobox is validate I can fix the axis in order to do a selection otherwise If it's not I've the possibility to make an operation of zoom analyzing the details. It's an idea but I don't know if it's work because I'm a beginner of this new library can some one could help me in develop it? Thanks a lot

interval = alt.selection_interval()
zoom = alt.selection_interval(bind='scales', encodings=['x'])
#slider = alt.binding_range(min=0, max=29, step=1, name='cutoff:')
#selector = alt.selection_single(name="VALUE", fields=['cutoff'],bind=slider, init={'cutoff': 0})
scatter=alt.Chart(pd_df).mark_line().encode(
    alt.X('@timestamp:T', title='TIMESTAMP'),
    alt.Y('VALUE:Q', title='VALUE'),   color=alt.condition(interval, 'VALUE:O', alt.value('lightgray')),
  tooltip = ['VALUE:O', alt.Tooltip('@timestamp:T', format = '%Y-%m-%d %H:%M')]
).add_selection(
    zoom,interval
).interactive(bind_y=False)

[jakevdp]

One thing you can do is set it so that one of the selections requires holding the shift key. There are some examples of this in the docs: Composing Multiple Selections

[LombardoAndrea195]

@jakevdp Thanks for the fast reply, I manage to respond now due to the fact before I didn't understand the logical of this implementation.(I've found out a step on github or stackoverflow where you did this issue for someone). I've another issue. I would like to represent in the time has before the number of events that occur in a day. I can make you an example suppose the example below:

input_dropdown = alt.binding_select(options=[list(pd_df_time['_SIDE'].unique())])
selection = alt.selection_single(fields=['_SIDE'], bind=input_dropdown, name='FIELD: ')


#shape
selection_SERIAL_NUMBER = alt.selection_multi(fields=['_SIDE :N'],bind='legend')
shape_SERIAL_NUMBER  = alt.condition(selection_SERIAL_NUMBER ,alt.Shape('_SIDE:N'), alt.value('lightblue'))

color = alt.condition(selection,alt.Color('count(@timeEvent):N'),alt.value('lightblue'))

interaction1 = alt.selection_interval(bind='scales', on="[mousedown[event.altKey], mouseup] > mousemove", translate="[mousedown[event.altKey], mouseup] > mousemove!",
    zoom="wheel![event.altKey]"
)

interactionY = alt.selection_interval(bind='scales', encodings=['x'],on="[mousedown[event.shiftKey], mouseup] > mousemove",translate="[mousedown[event.shiftKey], mouseup] > mousemove!",
    zoom="wheel![event.shiftKey]")
ScatterTime=alt.Chart(pd_df_time).mark_bar().encode(x=alt.X('yearmonthdate(@timestamp):T', title='TIMESTAMP'),
                                                               y=alt.Y('count(@timeEvent):Q', title='count()'), 
                                                               color=color,tooltip = ['hoursminutesseconds(@timeEvent):T',
                                                            alt.Tooltip('@timestamp:T', format = '%Y-%m-%d %H:%M'),'SERIAL NUMBER:N']
).add_selection(interaction1,interactionY,selection,selection_Operation).resolve_scale(  x='independent').transform_filter(selection)


ScatterTime.properties(width=1000)

For helping you to detect what I'm really searching about, I would like to visualize the number of times that in a day I've that event. I suppose that the logic could work but something doesn't because I've seen for example that in a day specific I've a count() of 2 but the serial number is one. I would like to inform you that the value is a time T but in this case I'm not interested on the content on the value but only if it is or not because I've defined a value time t that describe when finish the cycle of work. So more easily I want to evaluate the number of cycle in a day.

[joelostblom]

I'm closing this as it is already tracked in VL and there is nothing to do from the altair side of things. See https://github.com/vega/vega-lite/issues/1657