Ask Me Anything: Data Integration with FME

I still struggle with getting the schema back into the stream after reading from FeatureReader and getting a generic output.

I know the attributes are there b/c I can see them in the columns of the attribute viewer but I cannot dynamically expose. I asked Copilot and went on a fabulous hallucination filled trip of AttributeExploders Scanners, Mappers and ListExploders and FeatureJoiners :D 
 

With dynamic writing I get. that is smooth.  but how to read dynamic feature and get the column names back into the stream?

Another attempt example.
 

Looking forward to AMA! 

May be related to previous. Using feature reader on multiple files bring read from s pathname directory reader. And then exposing all attributes read from that dynamic source. Very difficult. 
 

second bump in the road is connecting updating, writing to enterprise arcgis still does not work for me. Perhaps the sso or some other network issue. Duo or other. 
 

And finally. I need one final upgrade for glow to get to where ssl/https/certs etc can be upgraded without assistance from safe. 
this would allow another question about data virtualization 

When resolving domains in an ESRI GDB Reader or an ESRI SDE Reader, FME creates new attributes with the _resolved suffix. To clean up the schema, I use a BulkAttributeRenamer to remove that suffix, effectively attempting to "overwrite" the original coded value attributes with the resolved string values.

However, I have experienced different behaviors depending on the writer:

1. Using a Classic GDB Writer, the translation fails due to schema validation/type conflicts.

2. Using a FeatureWriter (set to GDB), the translation succeeds without issue.

My questions are:

1. What are the fundamental underlying differences in how the Classic GDB Writer and the FeatureWriter handle schema mapping?
2. Why does the Classic Writer strictly enforce the original reader schema (causing failure), while the FeatureWriter appears more flexible with renamed attributes?

3. Is there any way to remove the ‘coded’ attributes and replace the _resolved attributes with the original name (example: STATUS_resolved renamed to STATUS (which originally has the coded values) but with the description values

Is there an updated method to replicate this: Question of the Week: Dynamic (but Hacky) Updates to Published Parameter Choices | Community in newer versions of FME?

Dynamic Updates to Published Parameters would be life changing!

As far as I am aware, there is no longer a ‘Feature Types to Read’ user parameter. I believe it got removed.

Therefore, is there another workaround to achieve this in FME 2025+?

Dynamic workspaces, as it is always a struggle to get them right 😉

SchemaScanner tips and tricks

SchemaMapper tips and tricks

We are aiming to implement a fully dynamic approach using a data‑mapping table that defines source tables and columns and their corresponding target tables and columns. However, so far we have not been able to achieve a truly dynamic method in FME to convert a source table into a target table where column names are systematically renamed from source to target.

We explored several approaches, including:

• ListBuilder combined with AttributeCreator and AttributeExploder
  This approach requires manual configuration, meaning attributes cannot be handled dynamically.

• AttributeExploder with row‑level source columns
  In this scenario, source columns are converted to rows and joined (via FeatureMerger) to the target column definitions. However, transposing the result back from rows to columns in a fully dynamic way is not supported by FME.

We also explored various suggestions provided by Copilot, but so far without success.

Ideally, we are looking for a method that supports fully dynamic mapping from source tables and columns to target tables and columns, driven entirely by a mapping table. We hope such a dynamic solution is available within FME.

Q1: We (local govt.) are thinking of traffic routing with FME.

When does FME integrate IDF = INTREST data format (whereas INTREST = Intermodal referencing system for traffic related data)? Or is there any possibility to connect with openrouteservice? or is there a commercial "routing service" by e.g. Google, TomTom or Here?

Q2: What is the most import characteristic (or talent?) one human must have to build useful workspaces aside from geospatial but in commercial/administrative fields?