Coincidentally, you might want to take a look at the post/blog from @mark2atsafe in which he talks about instead using AttrbuteCreator in "Adjacent Feature" mode.
https://knowledge.safe.com/questions/104425/question-of-the-week-variables-adjacent-features-a.html
What you could do is run the features through a Sorter (by Date).
Then run these through an AttributeCreator, but with Adjacent Features turned On. Number of Prior Features = 1.
Then you can use an Attribute Value in AttributeCreator that uses a formula like:
AccumulatedWastage=Â @Evaluate(@Value(featureÂ-1].AccumulatedWastage)+@Value(Wastage))
The only bizarre thing I found in FME2018 was that I needed to create "AccumulatedWastage" as an Attribute with a Default value first in a preceding AttributeCreator before I could run an AttributeCreator on this in Adajcent Feature Mode that updated its value with the Cumulative amount.
Try the Cumulative output port of the StatisticsCalculator, it will keep your running total per feature:
Coincidentally, you might want to take a look at the post/blog from @mark2atsafe in which he talks about instead using AttrbuteCreator in "Adjacent Feature" mode.
https://knowledge.safe.com/questions/104425/question-of-the-week-variables-adjacent-features-a.html
What you could do is run the features through a Sorter (by Date).
Then run these through an AttributeCreator, but with Adjacent Features turned On. Number of Prior Features = 1.
Then you can use an Attribute Value in AttributeCreator that uses a formula like:
AccumulatedWastage=Â @Evaluate(@Value(featureÂ-1].AccumulatedWastage)+@Value(Wastage))
The only bizarre thing I found in FME2018 was that I needed to create "AccumulatedWastage" as an Attribute with a Default value first in a preceding AttributeCreator before I could run an AttributeCreator on this in Adajcent Feature Mode that updated its value with the Cumulative amount.
This approach will be a lot faster than using the StatisticsCalculator
Coincidentally, you might want to take a look at the post/blog from @mark2atsafe in which he talks about instead using AttrbuteCreator in "Adjacent Feature" mode.
https://knowledge.safe.com/questions/104425/question-of-the-week-variables-adjacent-features-a.html
What you could do is run the features through a Sorter (by Date).
Then run these through an AttributeCreator, but with Adjacent Features turned On. Number of Prior Features = 1.
Then you can use an Attribute Value in AttributeCreator that uses a formula like:
AccumulatedWastage=Â @Evaluate(@Value(featureÂ-1].AccumulatedWastage)+@Value(Wastage))
The only bizarre thing I found in FME2018 was that I needed to create "AccumulatedWastage" as an Attribute with a Default value first in a preceding AttributeCreator before I could run an AttributeCreator on this in Adajcent Feature Mode that updated its value with the Cumulative amount.
There's an exercise demonstrating this in the current FME Desktop (advanced) training course: https://s3.amazonaws.com/gitbook/Desktop-Advanced-2019/DesktopAdvanced1Attributes/1.Exercise3.html
I think the part about creating the attribute might be resolved by using the Substitute Missing Values parameter (basically the first feature can't have featuren-1] because it doesn't exist, so it's necessary to have a substitute value).
Try the Cumulative output port of the StatisticsCalculator, it will keep your running total per feature:
Just be aware that there was a problem with this in various 2019.1 builds, where the cumulative port wasn't working. As long as you have the latest FME installed, you will be fine.
There's an exercise demonstrating this in the current FME Desktop (advanced) training course: https://s3.amazonaws.com/gitbook/Desktop-Advanced-2019/DesktopAdvanced1Attributes/1.Exercise3.html
I think the part about creating the attribute might be resolved by using the Substitute Missing Values parameter (basically the first feature can't have feature[-1] because it doesn't exist, so it's necessary to have a substitute value).
@mark2atsafe Finally did a bit more testing on this with FME2018.1.1.2 and there does indeed appear to be a small bug here. When enabling "Adjacent Features" mode on AttributeCreator, Workbench will say that the Transformer is "Invalid" if that Attribute has not been already created by a previous transformer ie. In Adjacent Features mode, Workbench will not validate if trying to simultaneously create and calculate the Attribute within the one AttributeCreator transformer. It seems to get confused with trying to validate the schema in thinking you can't update an Attribute that doesn't already exist from upstream transformer processes.
It appears to be a Workbench schema validation tester bug rather than a bug with the actual Transformer. If you blow through the warning of "running the workspace with invalid transformers" the AttributeCreator transformer in Adjacent Features mode will work just fine and both create and calculate the Attribute within the one Transformer!
Not sure if the bug has continued through to 2019+
You are right in that you do have to use a Substitute Default value to give a starting value for Feature at "Index -1", and the transformer is clever enough to create a temporary "dummy" feature before the first real feature with this default value, and then this starting value will get used for the real first Feature at Index 0.
Â
