T O P I C R E V I E W |
CalSchrotenboer |
Posted - 06 Dec 2012 : 07:29:13 If my data consists of 23 points and my grouping is set to 10, I would expect to plot three aggregate data points, two consisting of 10 individual data points and one of 3 individual data points. In practice, however, it appears that only two data points are plotted.
Am I correct in assuming that the last aggregation is ignored if it consists of fewer than the full complement of grouped individual data points. This appears to have the most significance in the case where the number of points to group is greater than half of the total points (eg grouping of 15 in the case of 23 data points) because then there remains only one aggregate point to plot and the charting routine throws an exception because it can't properly calculate the height and width of the area to chart. |
5 L A T E S T R E P L I E S (Newest First) |
quinncurtis |
Posted - 07 Dec 2012 : 16:36:47 It is not a valid X-Bar R Chart (using fixed sub group size) if you permit incomplete sample subgroups. See any of the reference books we refer to in the manual.
We cannot reproduce the problem you describe with missing the last data point. So start with one of our example programs and reproduce the problem using that. Then send a zip file of the complete project to support@quinn-curtis.com.
Just make sure you call RebuildChartUsingCurrentData after you add the last group of data, else the data will not show in the graph. |
CalSchrotenboer |
Posted - 07 Dec 2012 : 14:48:52 Thanks for your explanation. I will check with my supervisors regarding how they would like to handle this issue.
In our case, however, even though we call
spcChart.ChartData.AddNewSampleRecord(groupedDataPointsToPlot);
X number of times, only X-1 data points are plotted. We are using the following overload:
public int AddNewSampleRecord(DoubleArray samples);
When I set a break point, I can see that ChartData holds X data records in the CurrentNumberRecords property. Is there some other place where I can look to see why the last record is being dropped. |
quinncurtis |
Posted - 06 Dec 2012 : 09:48:58 What you are describing is invalid. You cannot use a regular x-bar r chart, or x-bar sigma chart, and on the last sample interval, suddenly decide to change the number of samples in a sample subgroup (changing from 10 to 3). You must either have a full sample subgroup, or not include it. Anything else will produce invalid results based on the formulas used in calculations for the charts. However, in our simple tests, when we did update the last of 3 sample intervals (with fixed subgroup size 10) with an array sized to 3, we still see three data points, one for each sample interval update. Though we know that any control limit testing against that point will be invalid.
So, I don't know where you want to go with this, since your underlying premise is incorrect. |
CalSchrotenboer |
Posted - 06 Dec 2012 : 09:09:34 Chart Type: X-Bar R (or X-Bar Sigma)
number of samples per subgroup = what I was referring to by grouping
Trying again:
If my data consist of 23 points and my number of samples per subgroup is set to 10, I would expect to plot three batches, two consisting of subgroups of 10 data points each and one with a subgroup of 3 data points. However, even though in my code AddNewSampleRecord is called three times, only two batches are plotted.
Am I correct in assuming that the last batch is ignored if it contains fewer than the full "number of samples in subgroup". This appears to have the most significance in the case where the number of samples in subgroup is greater than half of the total number of data points (eg a samples in subgroup size of 15 in the case of 23 data points) because then there remains only one batch to plot and the charting routine will throw an exception because it can't properly calculate the height and width of the area to chart.
|
quinncurtis |
Posted - 06 Dec 2012 : 08:51:11 Sorry, but I do not understand this description. I don't understand what you mean by aggregation of data points. That is not terminology we use anywhere in the software Please explain it again. This time, refer to the exact chart type, the number of samples per subgroup, and the number of batches (or sample intervals) you have, and how many time you call AddNewSampleRecord with an array of data sized to the number of samples per subgroup.
|
|
|