Apache OpenOffice (AOO) Bugzilla – Issue 94632
behavior of 'first row/column as label' when given ranges are complex
Last modified: 2013-02-24 21:20:35 UTC
creating a xy-chart of the data area e.g. $Tabelle1.$A$1:$B$9 and wish to not display the first data in A2:B2 in the speasheet in the chart using the second tab of data area results in the data area: $Tabelle1.$A$3:$B$9;$Tabelle1.$B$1 o.k. The defect: changing the data area in the first tab to e.g. "$Tabelle1.$A$4:$B$9;$Tabelle1.$B$1" yields to the displayed data (see second tab): name: $Tabelle1.$B$1 x: $Tabelle1.$A$5:$B$9 y: $Tabelle1.$B$4:$B$9. There is always the value "ONE" added to the first cell for the x-values, independent if the user changes the first or the cell of the data area. this defect is in OO 3 RC1 to 3, I have not checked OO 2.4.1 because I need the displayed regression function for some hundreds of data points.
Created attachment 57005 [details] simple x-y-test chart
@IHA: i think the behavior is correct but not was the user want.
sorry, the behavior is not correct, it is not only that an false range is displayed, but also the x,y-pairs are shifted: in my example the first displayed pair of data are the cells A5:B4 and so on. 6 data point should be displayed, but there are only five! The second signal is, that the regression function became false, it changed from 2 x 2^x to 2^x. But I have to refine: when working without "first line as header" checked, the behavior is correct. You can see it, observing the graph and the marking of the cell range. With "first line as header" checked the first data cell in the A-column is marked separately, also after manual correction of the data range.
The 'First row as label' check box works as follows when it is checked and data series are selected to be in columns: The whole range is split into columns. Then the first row of each column is taken apart from the rest of the column to be used as potential label. In that sense the behavior is correct. You have entered an asymmetric range. That is the reason why you get an asymmetric result. When you enter a symmetric range for example $Tabelle1.$A$1:$B$1;$Tabelle1.$A$4:$B$9 the result will be what you expect. In general when you enter the range on the 'Data Range'-page of the wizard then there are only these simple splitting advises which do not lead to the wanted results under all circumstances. For more advanced splitting of the given data ranges we introduced the page 'Data Series'. Here you have the free choice to set all ranges for values and labels as you like.
The asymmetric range, regarding to the "first line as header" is generated by the chart module not by the user! The reason is the selection of data to be displayed (e.g.in different charts). Above I used first / second tab instead of the 'Data Range''Data Series'. In my practise there are data sheets with some thousand data in continous rows which have to be displayed in different charts, according to changed working parameters. When the data rows are side by side its the convienient way to change ("fine tune") the displayed range in 'data range'. A workaround is, as you described, to reintroduce manually the whole cell range for the "header". Then the changing of the data selection generates no error. A shorter way to get the "asymmetric range" is to import analogous worksheets with charts from excel.
Sorry, I do not get your point. The 'Data Range' page is intended for use with simple rectangular ranges. You are a power user with more complicate demand for the range splitting. For power users we have introduced the page 'Data Series' where you can define more complicated ranges manually. Those more complicated splitting simply cannot be displayed on the 'Data Range' page correctly in all cases as that is beyond the capability of that page. Still when you insist on using the 'Data Range' page I have given you a workaround to do so for your cases. So is there a problem still?
I close this issue now as the problem is unclear.
closed