Skip to main content

Hi,

i have an issue with the statistic calculator, doing some IDW Interpolation.

I use a counter (the only one within the workbench) to create an id, called "_zaehler". Later on I use this attribute as a group by attribute in the statistic calculator. The point is that, with the illustrated settings (count start=3000), I loose some points (with theses settings 530 points!). The missing points look like this:

If I set the count start to, say 10.000, all points are processed! This behaviour is unclear to me!

If you use a duplicate filter instead of the statistics calculator, do you get 20,703 featuers out the unique port or 20,173?


Hi @egomm,

pls. refer to the attached screenshots. There are excatly 20703 unique values.

I found a related thread where a "string formatter" is recommended before the statistic calculator.

https://knowledge.safe.com/questions/54930/statisticscalculator-bug.html

This solved the problem. Nevertheless I guess this is a bug.


Hi @egomm,

pls. refer to the attached screenshots. There are excatly 20703 unique values.

I found a related thread where a "string formatter" is recommended before the statistic calculator.

https://knowledge.safe.com/questions/54930/statisticscalculator-bug.html

This solved the problem. Nevertheless I guess this is a bug.

Thanks for bringing this to our attention @femo. This is a known issue related to the post you linked, and our development team is on it. Check back at https://knowledge.safe.com/questions/54930/statisticscalculator-bug.html for updates.

 


Hi @femo , what build of FME Desktop are you using? You can find this by going to Help > About FME Workbench.

Also would it be possible to share a version of your workspace so I can add it to our PR so the developers can look at this more closely.

Thanks,

-Liz


Fix is going in now. See details over at https://knowledge.safe.com/questions/54930/statisticscalculator-bug.html


Great news @femo, the bug has been fixed in the latest hot fix 2017.1.1.1, which you can download at safe.com/downloads, it is also fixed in any future versions. Thanks again for catching this!


Reply