Newly-created plots with too many points to actually plot appear to have downsampling enabled even when it's not supported (i.e. when the X axis is not a datetime). For example, the following query:
t = emptyTable(250_000).updateView("RowIdx = ii").update("Val = Math.random()")
plot = figure().plot("Plot", t, "RowIdx", "Val").show()Creates this plot (notice the 'downsampling failed' error icon):
There is nothing indicating that downsampling should be used or would be required; the only thing you see is a blank chart and that downsampling failed.
When disabling downsampling, the UI isn't always responsive — the panel doesn't 'spin' or anything, so it's hard to tell that it's working. For example, when I click to disable downsampling on that plot, this is still all I see ~3 seconds later:
It does eventually draw (although for larger tables w/ millions of rows, it appears to hang indefinitely):
When downsampling or plotting fails, it would be helpful to display the number of source table rows — makes it easier as a user to understand what you're telling DH to do/what you're doing wrong.
Newly-created plots with too many points to actually plot appear to have downsampling enabled even when it's not supported (i.e. when the X axis is not a datetime). For example, the following query:
Creates this plot (notice the 'downsampling failed' error icon):
There is nothing indicating that downsampling should be used or would be required; the only thing you see is a blank chart and that downsampling failed.
When disabling downsampling, the UI isn't always responsive — the panel doesn't 'spin' or anything, so it's hard to tell that it's working. For example, when I click to disable downsampling on that plot, this is still all I see ~3 seconds later:
It does eventually draw (although for larger tables w/ millions of rows, it appears to hang indefinitely):
When downsampling or plotting fails, it would be helpful to display the number of source table rows — makes it easier as a user to understand what you're telling DH to do/what you're doing wrong.