Unlock a world of possibilities! Login now and discover the exclusive benefits awaiting you.
Hello,
I was hoping someone could explain to me how the dimensions work in a Chart (or in general, but I am currently working on a chart).
I have 2 confusion points, firstly it seems the order that the dimensions are in on the "Used Dimensions" section impact what is displayed on the X & Y axis... but I can have as many dimensions as I want... how do all the extra dimensions get taken into consideration? Is there a guide how this works (I can't find one)? Right now it seems my expression creates the Y axis and the first dimension is displayed on the X-axis, but doesn't seem consistent... read next part please...
Second, on my specific chart, I have 3 dimensions, let's call them 1, 2 and 3 (in order). I noticed when I apply a filter to Dimension 1 and the filter is down to a single result, Dimension 2 now appears on the X axis instead. I do not understand this logic or why it does this. I can understand that it might not make sense to keep Dimension 1 on the X-axis, but how can I take advantage (or what exactly is it trying to do) so I can have it display a chart that makes sense? Changing Dimension 3 to dimension 2 spot does not produce a graph I am happy with either, but I'm not sure how to get it to do something I would prefer without adding more dimensions which then impacts the larger view of the chart (i.e less filters).
Thanks for any explanation.
The online help for Chart Types / Chart Properties Dimensions: contains the following guidance on number of Dimensions per chart Type.
"A chart can display one or more dimensions. The upper limit depends on the actual chart type, the complexity of the data and the available memory. Pie, line and scatter charts can display a maximum of two dimensions, bar, block and grid charts three. Radar and funnel charts can only display a single dimension, and gauge charts use no dimensions at all. Further dimensions are ignored."
https://help.qlik.com/en-US/qlikview/November2017/Subsystems/Client/Content/Bar_Chart.htm
The Doc says "Further Dimensions are ignored" which is true when there are multiple values for the upper Dimensions. Line charts can use a maximum of 2 dimensions so the third dimension is ignored and has no effect on the rendering of the chart. Dim1 is used as the x-axis and Dim2 generates multiple lines for Dim2 values. Until...there is only one value for Dim1 one and then Dim2 becomes the x-axis and Dim3 generates multiple lines.
I'll admit this dynamic promotion/drill behavior is entirely new to me and I've been using QV for 12 years. Of course I only ever used 2 dimensions on my line chart because I "knew the dimensions beyond 2 were ignored". Hah.
Perhaps someone else on the forum has some insight as to whether this is an intentional feature or a bug.
To offer some relaxation about your original question, there is documentation about how many dimensions are allowed and what they do. Yes, they are order dependent. In the case of a line chart, Dim1 being the x-axis and Dim2 generating individual lines. What's surprising to both me and you is the automatic promotion of Dims when there is only one value.
-Rob
Thanks for the reply, however this is one of the articles I did find, however he doesn't really go into specifics of how the orders of the dimensions impacts what is displayed on the chart, unless I'm missing it.
Hi,
a good example...
2-dimension bar chart with labels of both dimensions displayed at x-axis
HTH
André Gomes
hello,
Please be more specific about your request: what kind of chart are you using? what kind of mesures or aggregation?
please note that all the chart do not use the same number of dimisions.
an exmple should be very helpfull !!
regards,
I am currently working on a Line Chart, however I just would like to understand why solutions like the one above even impact anything that is displayed.
Why are people adding "dummy dimensions", what is it about QlikView that it even needs one? There are a number of examples having that as a solution, but none of them really talk about why it is even necessary. Seems kind of silly to me to need to add a "dummy dimension", but I figure it is probably because I don't get how QlikView is interpreting that dummy dimension.
Also why when I filter to a single result on the top dimension, does it change the x axis dimension that is used?
This reply and the one above both show how to do something, but no explanation as to why they're doing that. I like to know why and not just how!
Thanks for responses.
Your behavior of a single value drilling down to the next dimension sounds like you are using a Dimension group instead of three individual dimensions. Is that correct? If so, this is the expected behavior for a Drill-down Group.
What type of chart are you using?
-Rob
I have 3 dimensions pulled straight from my data, one at a time. No calculated dimensions / dimensions group (what is this?). It is a Line Chart. When I sort them in a different order (promote / demote) I get different results, and then further each result could be different if I have any filters applied, as in they COULD look different per filter (although not always the case).
It seems like QlikView very much cares about the order of these dimensions and how they work with the selected filters. Unfortunately I can't post any of this data , but basically I have Server names, then database names, and then a timestamp. My expression calculates a percentage based on a Boolean value, but basically its a number greater than 0%, less than 100%
I have the Timestamp as my x-axis (so I list it first in the dimensions), then server name and then database name... if I change the order of Server / database, my resulting chart looks different... when I put server name then database, i get an up and down trend just as I expect... as soon as I put servername first, it's just a flat line at 100% with no filters applied. When I leave it as server name and then database, but then apply a single timestamp as a filter, the x-axis turns into Server name instead of leaving it as a timestamp.
It all just seems pretty random, I just keep guessing because there is no real guide as to what adding a 3rd, 4th, 5th etc dimension actually does or how QlikView prioritizes them (at least no guide that I can find) and how it impacts what gets displayed on the chart.