Are there known problems with the Sorter transformer in FME Workbench version 2021.1?

Did you choose numeric in the sorter?

Likewise experienced this in FME 2021.

The same transformer versions and parameters worked fine in fme 2019 so this would be a bug in 2021.

And yes, using the "Numeric" option.

My workaround was to put an ExpressionEvaluator beforehand to create an _order.hundredths attribute: @floor(@Value(_order)/100)

and then to put this in the Sorter:

• _order.hundredths | Numeric | Ascending
• _order | Numeric | Ascending

Hi @geopositie​ I haven't been able to reproduce this issue in FME Desktop 2021.1 with a few quick tests. If you download the latest version of FME Desktop 2022 and test the same workspace (after updating all Readers and Transformers) do you still run into the same issue?

If so would you be able to share a sample workspace & dataset with us here that reproduces the issue?

I noticed an issue today that may or may not be related, when using 3 levels of sorting. Specifically, the sorter fails to correctly sort the 3rd level when the string is complex, in my case includes parenthesis, e.g., "BAFX39C (1-32)". Strangely, when use a single level of sorting, it does work correctly. Maybe a memory issue? Anyway, I created a new parameter dropping the "(1-32)" and used that new parameter for the 3rd level of sorting and it worked fine. Incidentally I am using the "Natural" sort setting typically, but in this case, I tried "Alphabetic" and "Numeric" and that did not solve the problem.

@pmcpherson​ are you able to share a copy of your workspace and data or a screenshot of the data and the Sorter parameters? I would expect FME to still be ok with sorting through three levels, but as you mentioned something in the logic could be causing the Sorter to fail.

Additionally, what version & build of FME Workbench are you using?

I'm working with proprietary data, so no, I can't really share the data or workspace. I'll try to reproduce the problem with fabricated data and will share if that also exhibits the problem.

@danminneyatsaf​ I set up a "dummy" routine today and was unable to reproduce the problem. I then reran the (proprietary) routine from yesterday and it also did not fail, even though nothing had been changed. Odd.

That's very odd. I've tried reproducing the issue myself in FME 2022.2 as well and haven't had any success so far. If you encounter the issue again, then please let me know and we might need to move over to a support case to better investigate the issue.

Sounds good @danminneyatsaf​ . By the way, you had asked which version I am running. It is the latest GA, 2022.2 Build 22765. Thanks again.