The problem I'm having is my cumulative percentage line is not correctly summing over subsequent X categories that have the same value (see attached file.

[Nick Zubillaga]

I created a dynamicPareto chart.The problem I'm having is my cumulative percentage line is not correctly summing over subsequent X categories that have the same value (see attached file.)

 

Count([Order]) THEN Sum([Value]) OVER (AllNext([Axis.X])) / Sum([Value]) OVER (All([Axis.X])) AS [Cumulative Percentage]

 

Any thoughts or suggestions

[Manoj Chaurasia]

Thank you for the response. However, that option is not available (greyed out) in the properties box.

I find it odd that my cumulative line calculates properly for all x-axis categories except for 6-7-8 which are each count=2 and 9-10-11 which are each count=1. The cumulative line calculates correctly only for the first occurrence of redundant values, i.e. at category 6 and 9.

[Tirupathi Mangalarapu]

Hi Nick, Can you please select the check box "Compensate for missing value" when you right click on visualization properties >> Appearance Tab. Let me know if that helps(I have also attached a screenshot for your reference)s

[Nick Zubillaga]

It's acting as if the (AllNext([Axis.X])) is seeing all next values and not all next categories. (I.E., category 6, whose value is count=2, has the cumulative percentage for ALL categories (6,7,8) whose count=2)

[usha p]

Hi, I'm having same problem. The Cumulative percentage does not work when there is same values encountered. Can anyone suggest a solution for this. Thank you.

[Brad Car]

I have a solution, although not ideal in my opinion (neither is the whole pareto chart implementation in spotfire, but thats a seperate issue). I created 2 columns, P1 and P2 (both just used for the Pareto chart). For my data set, I was doing a count of rows. If this is for a sum of values, you would likely need to modify this process slightly (may not work)

 

P1 is a random value, the exact value doesn't really matter but it should be somewhat large. For my example, I used RandBetween(1000000,2000000,8)

 

P2 is the value I count over, and is basically a real very close to 1, but not 1 such that I know longer have repeated category values, so (1 + (([P2] - 1) / [P2])) / 2. This number is very close to 1, and random.

 

I then set my x category by this, and no longer have any ties (since the values are fractionally different). This solution is not ideal because you get a long string of numbers along the bottom (the current solution has this now but are integers)

 

 

[thijs laag]

I've found an actual "solution" to this problem (or at least found what is causing it, cannot reproduce the faulty behavior very well). The trick lies in the hierarchy on the x-axis. For some reason the hierarchy is missing when I followed the instructions from the youtube movie TIBCO Spotfire - Combination Chart Configured as a Pareto Chart | TIBCO

Software. See first picture:

When I tried all steps in different order another time, somehow it grouped categories on the x-axis with the same value as an hierarchy; and thereafter the cumulative percentage line was displaying correctly :) Don't know exactly what is causing it, but at least it works now...