Apache OpenOffice (AOO) Bugzilla – Issue 81947
Missing points in XY chart
Last modified: 2013-02-24 21:21:55 UTC
It was tested in Windows (currently I don't have Linux version of OO.o 2.3). I will attach a file illustrating the issue. In this file "chart_bug.ods" in the Sheet3 there are data in the first two columns these data are plotted in the chart. The central part is completely missing. The same plot in OO.o 2.2 is correct.
Created attachment 48489 [details] bug in creating a cgart in OO.o 2.3
Created attachment 48490 [details] bug in creating a chart in OO.o 2.3
wrong component, changed to Chart and re-assigned
It happens too, if the chart is newly created in OOo2.3. It doesn't depend on logarithmic scaling. You can notice it too, if you switch to normal scale. You see it, when you traverse the points with <tab> and <shift><tab>. It is not the amount of data points, I'll attach a smaller example.
Created attachment 48492 [details] missing points in newly created chart
Too avoid the creation of too many data point symbols for large charts, some data points may be removed (if they lie on the same point in the drawing layer coordinate system). That's why the number of data points is less than expected in the non-logarithmic case. There, I would see the behavior as correct. For the logarithmic case it looks very much like this algorithm would not take the scaling into account, i.e. the points have the same pixel position in the untransformed system, are then removed and then the scaling transformation is applied. This is a bug. ->iha: Please have a look.
changed component to Chart as it is clearly a Chart and not a Spreadsheet issue. Frank
and back to Ingrid. Frank
sometimes I'm to fast.
set target
Fixed in CWS chart18.
Created attachment 48522 [details] patch to fix the bug based on oog680m5
the attached patch has an error, I'll attach the corrected patch
Created attachment 48527 [details] corrected patch
->Thomas, please verify in CWS chart18.
seen ok in cws chart 18 -> verified
*** Issue 82912 has been marked as a duplicate of this issue. ***
seen ok in current master -> closed