Apache OpenOffice (AOO) Bugzilla – Issue 52845
Use time scale for x-axis in scatter chart
Last modified: 2013-02-24 21:19:42 UTC
I am trying to solve a simple issue with Excel, Lotus 1-2-3 and Calc since years and see no chance to do so as Microshit invented charts linked to spreadsheets coming (as usual) from the wrong side and by setting the industry standard dragging all competitors into the same dead end street. (What angers me most is the fact that in the times of an IBM 360 or S/3 in the 60ies and 70ies the way to handle this was there - easy and clear.) Try the following simple scenario: You are running/biking or move in whatever way irregularly on a training course and note the total time. You enter the date and running time in a Calc row using one spread sheet page within one year or season. (If you only use minutes and seconds you will encounter in Calc some completely unnecessary mistakes and inconveniences because MS didn't separate functionality and user interface; I reported this issue to OO some time ago but you cannot correct it because you have to remain compatible with MS) Now you want to create a chart and chose an xy-type and lines with symbols. The time is on the y-axis, the x-axis is defined by the date differences of the running dates (in days; can easily be calculated in additional columns). How can you now correct or adjust the following flaws: - on the x-axis you do not want a grid with days but for every entry the date printed, (although the axis is based on a column in your Calc sheet). If printing space is a problem you would need a mark on the base line of the x- axis with a number or color to link to a legend below. Eventually you would like to have a vertical line up to the time point for every date. - you might like to define the minimum, maximum and interval of either grid yourself. How can you in the course of time expand, enhance and correct your chart or combine it with other charts that have the same axis definitions and at least similar grid values without every time recreating it? (e.g. combine different seasons or compare to a friend that does the same training) etc, etc Imagine you would start with calling a chart program, where all these details mentioned can be defined at your will and in the last steps or in any additional ones later you define for each axis, and line (in this case) the underlaying Calc row or column or single value. Wouldn't that be flexible and easy to handle. Wouldn't that open all possibilities of imaginable combinations without the nightmare of recreating everything everytime from scratch? Wouldn't that allow to define some charts as templates (like the ones you supply now) with the specific scales for your purposes and in the last step define the Calc values. Or just reuse an existing chart exchanging the Calc values. The most flexible part of a chart is the values it shows; the way these charts are built now (and the mess it creates in practical life) is as if a cook would cook the meal before chosing the pan! With best regards Bruno Naumann
As far as I understood your problems, you want to have the difference of the current running date with the previous one as x-values, to get a chart in which you can later read "my times are pretty bad if I didn't run for two weeks, but if a run every other day the time is quite good". Did I understand you correctly so far? You want every date printed, but where? On the axis there is usually only one label per tick, do you want a feature to put more than one label at a tick? Or do you want the dates in the chart area -- that is possible via Insert/Data Labels. Btw., why do you need the actual dates in the chart? Do you want to have the additional information about the time of the year, like "in winter greater breaks are worse than in summer"? Minimum, Maximum and Interval of grids can be set via the Scale tab-page of the axis property dialogs (the axis scaling is identical to the grid scaling).
... and furthermore: what do you mean by the Summary? Please set the summary to a short ddescription of your problem, otherwise it is unclear what this issue is about when being in a list of many issues.
Before I answer some of the detailed questions of yours let me put the issue in the overall context. This example is just a hobby of mine and does not need to be solved with whatever circumventions and tricks. But all the same I think it is in its kind something that could be very handy in business - but only if you tackle the problem from the opposite side as I outlined in my block beginning with "Imagine ...". (E.g. compare the effect and timeshift of promotion campaigns with sales results, compare the effect of promotion channels depending on the type of goods; etc etc). So it is in the first place an attempt to make you think about creating a graph module (standalone or as a part of your drawing application) that can be freely linked to Calc values as I described it. Context details: I could have based on my numbers a lot of additional comparison possibilities. E.g. whether the curves are similar to each other (over different seasons) or if the pattern is changing; the effect of the proximity of the runs on the result (can also be negative if the recovery time is too short) or just the speed of decline due to my aging. As I also have intermediate times stored in my Calcs (e.g. for a flat piece and for an ascending one of the track) there are many more comparisons possible. Details: the x-axis reflecting the dates in the year (or season) cannot be based on the actual dates but on the (additional column of) days elapsed since the beginning of the year (or season). So I get the correct spacing for every entry of running time in the appropriate column. But the text of the scale are these days the axis is based on (in a regular fashion: 20,40,60 etc) and not the dates - and these days have no meaning to a user. My idea that the actual dates of the table entries should be marked on the 0-line of the x-axis is just an ideal which can lead to unsolvable space problems especially when you combine different tables. What would come near to that ideal was an equally regular scale in a date format. But as the actual date column cannot be used as a useful axis base, a retranslation of our elapsed days in actual dates would even be more complicated. Facit: it could be a handy additional function to use calendar dates as a base for an axis to give the spacing. Thanks and best regards Bruno Naumann
As far as I understood this issue, it is about having the possibility to use date values for scaling axes. This is also the topic of Issue 25706. *** This issue has been marked as a duplicate of 25706 ***
closed as duplicate