Created attachment 138945 [details] Example screenshot When data is on a "wrapped" table (i.e., labels and their values are placed o two or more lines or columns -- see attached image and/or ODS file), the resulting graph (e.g., columns) will: a) show values correctly as separate categories, but b) place labels stacked accordingly to which column (or line) the belong (see ODS file). I would have expected the data in the labels field (as well as elsewhere) to be treated as a single series in the given order, regardless of how each individual cell maps on the spreadsheet (even if they are taken from separated sheets!). At least, when introduced or customized manually in the Data Series dialog window. Is this related with hierarchical axis labels? I understand that this occurs only when the first two or more columns or lines contain text (as per Calc Guide 4.1, Fig. 77 https://wiki.documentfoundation.org/images/4/47/CG41-CalcGuideLO.pdf).
Created attachment 138946 [details] example file
Repro 6.2+ as reported. Can you please explain how you created it to get data range: $Hoja1.$A$2:$C$3;$Hoja1.$A$4:$C$5;$Hoja1.$A$1 I'd says IMHO this is not "normal" bug in terms of wrong behavior because Calc doesn't support table prepared like that. Basic options for "Data series" are "in rows" and "in columns". What you ask for is an enhancement to get custom data series. I change the title. Should it be accepted is a question. My personal stance is no. I don't know whether UX or Dev should review this one.
The example originally attached was created only to show the noted behaviour. As you say, LO seems to be restricted to "basic" (single) rows and columns. Having to plot data sets that "wrap" over more than one row or column is maybe not as exotic, though. I remember having done so on MS Excel over 20 years ago (for my grad thesis). I have seen many other instances (e.g., accounting or budgeting) in which a data set may be separated in several tables (different sheets or not, depending on the user/company presentation/printing needs), but plots needed both for individual table data and for the whole lot. It wouldn't make sense to duplicate and re-arrange the data only for one plot. I do not have one of those with me now but I'll try to get my hands on a real-life example to see if I can share it.
Please describe shortly how you create the chart; I guess per manipulation of the data range (y-values = $Hoja2.$B$2:$B$4,$Hoja2.$D$2:$D$4).
No feedback, so let's take Timur's opinion and resolve as NAB. Workaround for this particular case is to add a dummy cell. For example $Hoja1.$A$2;$Hoja1.$B$2;$Hoja1.$C$2;$Hoja1.$A$4;$Hoja1.$B$4;$Hoja1.$C$4;$Hoja1.$C$5