https://bugs.kde.org/show_bug.cgi?id=393215
--- Comment #6 from urcind...@gmail.com --- (In reply to Alexander Semke from comment #5) > (In reply to urcindalo from comment #4) > > > > Yes, I already checked that. However, in 1. above, if a user is interested > > in plotting just one y-column, I find it intuitive to *only* select this > > particular column and expect LabPlot to automatically select its designated > > x counterpart, not itself as the x-axis. > Ok, I see. But how can we figure out what is its x counterpart? User can > have in principle an arbitrary mixture of plot designations for different > columns in the spreadsheet maintained. And the only thing that we have at > the moment is the plot designation for the column itself, if provided by the > user at all. For example, if the user has the following columns > > x y y x > > and selects the third column for the plot, what is its x counter part? > > > > > What's more. Imagine you wanted to plot a y-column from one spreedsheet vs. > > an x-column from *a different* spreedsheet. I think LabPlot should show the > > designated x-columns from the available spreedsheets in the drop-down menu > > when you just select one y-column and right-click on it to plot the data. > > Or, at the very list, as I said before, its corresponding correct x-colun > > instead of itself. > With plotting of data via the context menu of the spreadsheet you have an > easy tool to quickly plot the data, but for this particular spreadsheet > only. If we would show also other spreadsheets in this dialog, this will > cause confusions in many cases. If you want to plot data from different > spreadsheets you need to manually create a plot, add a xy-curve and provide > the column data sources. > > Also similar to the arguments above, we don't have the information about > "its corresponding x or y column", that's why we cannot do any magic here. > Even if we would allow the user to enter for every column the additional > information about the corresposding x/y columns, the user has to maintain > this which will result in additional overhead. I see. Maybe I'm too spoiled by QtiPlot :) In QtiPlot you know exactly which x corresponds to a y because it works this way: a) By default, the first column in any spreedsheet is marked as "x" and the second as "y". Then, every column you add is marked by default as "y", so its corresponding "x" is still the first one in the spreedsheet. No confusion. b) But, let's say the fifth column is marked also as "x". Then, automatically, the first column is marked as "x1" and all the columns to its right as "y1" until the 5th one, which is now "x2" and all the additional columns you might add to its right will be "y2". In the end, you just have to select the "y" column you want to plot and the programs selects its "x" accordingly. It's very convenient. Maybe this idea could be implemented in LabPlot in a future? Nevertheless, I still think the current behavior is a bug, because if you only have one spreedsheet with only two columns, one clearly marked as "x" and the other as "y" in it, even if you just select the "y" column, right-click on it and select plot, the column marked as "x" in the spreedsheet should appear as an option for the x axis in the corresponding drop-down menu, as is the case when the user selects "Plot data" from the context menu of the spreadsheet without previously selecting any column. -- You are receiving this mail because: You are watching all bug changes.
