Transformers

Good day,I have a table with a couple of records.I would like to keep one record per category and this record should be the record which has the less attributes with a null (missing) value (I highlighted them in green on the screenshot.). How can I automatically calculate which record per category I should keep? 

Hi all.I just experimented with a small workflow, that scans a folder for files, and attempts to read those files using the generic format.I tried with CSV and DGN, and both worked well.I then tried with Geopackage, and then it blew up.First because FME’s guess using the extension was wrong, so I had to parametricize the format used.Then it blew up with an internal error message (“Too many recursive macro substitutions”):2026-03-03 15:14:24| 2.6| 0.0|WARN |UniversalReader -- readSchema resulted in 0 schema features being returned2026-03-03 15:14:24| 2.6| 0.0|ERROR |Too many recursive macro substitutions -- line currently is: `SOURCE_READER GENERIC REAL_FORMAT "$(REAL_FORMAT)" EXPOSED_ATTRS "$($(FORMAT_SHORT_NAME)_EXPOSE_FORMAT_ATTRS)"'2026-03-03 15:14:24| 2.6| 0.0|ERROR |class stk::ex::UString<0> __cdecl tabrdr::ClassicTableReader::macroSu

​Hi everybody, I have a similar issue to the one the creator of this thread has: The difference is that I'm dealing with donuts. The workaround from Takashi (buffer issue location → dissolve → aggregate) doesn’t work with donuts. The reason why this doesn't work is explained here: The GeometryValidator reports self-intersections in the wrong place for donuts. If I buffer the issue location, part of the donut that doesn’t need to be buffered gets buffered, and the actual issue location remains a problem. I also tried the GeometryValidator repair function, but this didn’t work either. 

I have a donut geometry (see WKT below) which I am trying to validate. Using the GeometryValidator only looking for self intersections I get an error on the start/end point of the outer ring: Error report:Coordinates from the inspector: I would expect this to not be reported as a self intersection since polygons normally require intersection at exactly that point.  Why is FME reporting this? I have tried using both FME Desktop 2021.2 and 2022 beta. Geometry WKT: POLYGON Z ((558000 6142955.0838 0,558000 6142917.0404 0,557997.11 614

Hello everyone,I am seeking confirmation on whether GeometryValidator supports 3D geometry validation (such as solids, meshes, point clouds, etc.) in version 2025.0I noticed that GeometryValidator includes an issue type called “Self-Intersection in 2D”, which suggests that this one specifically built for 2D geometry. However, I would like to understand whether the other issue types - such as Missing Vertex Normals, Missing Texture Coordinates, Invalid Solid Boundaries, Invalid Solid Voids, Contains Non-Planar Surfaces, Invalid Area Orientation, Surface Orientation, and others - are applicable to and capable of validating 3D data.Could you please clarify which of these validations support 3D geometries?Thanks in advanced!

Hello I have FME 2022.2 version and i also could upgrade it to 2025 if it helps but i would like to extract data  Input data : *.pts or *.e57 file format (about 2 to 5 room point cloud data in One *.pts file/*.e57file)Desired output data : *.pts file format (ONLY 1 room to be extracted in *.pts file format)Other supporting data : Sitewide room model in *.ifc format (federated model having about 20+ rooms in 1 IFC format) TASK : Using IFC room model for any give Room for example (Room1), i would like to clip the input data set i.e. in *.pts/*.e57 file format and out only that particular room  (Room1) *.pts format data.  Am not sure how to set / which reader and writer / transformer to use to get it right . I understand i will have to use Clipper to clip *.pts file using IFC file somewhere. Highly appreciate and obliged with workflow for your support and know-how. 

Hi all, I regularly setup workspaces with quite a few feature readers & writers that need a lot of parameters changed when moving between environments. I make use of scripted parameters to make this easy based on an initial Environment selection.The one parameter I can’t get to update correctly with a scripted parameter is the web connection, does anyone know if this is possible?

Has anyone else experienced this? I'm using FME 2022.2. If I pick an attribute from the arrow drop down menu, it inserts it in the DateTime Attributes field as @Value(CREATETIMESTAMP). When the transformer runs, it doesn't convert the value. But if I select the attribute from the 3 dots menu, it's fineIn the 3 dots menu chooser, there are sort of two sets of the attributes....

Hi all,I am experiencing an issue with dataset fanout in my workbench after upgrading to 2025.2.0.0 with Build 25810 (WIN64). The workbench worked perfectly in 2024.2.The workbench reads in a FGDB file 30,000,000 features using a reader, then performs spatial filtering with coordinate grid tiles and fanout into geopackage files according to the tile names using a writer. The workbench always stopped while outputting the first geopackage file with the following error message.  Insufficient memory available -- error code was 2 - please read the FME Help section 'Adjusting Memory Resources' for workarounds.The workbench worked fine when I set the max features to read in the reader as 100,000, but the writer failed once the reader read all 30,000,000 features. I tried lowering the number of features per transaction in the writer but it did not work. Is anyone else experiencing a similar memory issue with dataset fanout in 2025.

Hello everyone,I have a workspace app that converts a .dgn file to a .gpkg An user tied to run the tool with a dgn file, but he received a failure message. I looked at the job and I saw the following errors 464 OGCGEOPACKAGE writer: Invalid curve: not enough points 754 OGCGEOPACKAGE writer: Invalid curve: not enough points 872 OGCGEOPACKAGE writer: Invalid curve: not enough points 957 OGCGEOPACKAGE writer: Invalid curve: not enough points 1051 OGCGEOPACKAGE writer: Invalid curve: not enough points    Did anyone have similar errors and what can I do to solve these errors?

There are a series of similar posts but still not seeing a way to expose attributes dynamically.I have a a large number of geodatabases which have multiple layers and different attributes, which I am merging with a geometry layer so I end up with a single layers with lots of attributes.The geodatabases have a series of tables and a geometry layer, but there is no consistency in attributes between the geodatabases.How do I expose and have these exposed attributes output to a merged dataset? 

I believe there’s an issue with the tolerance values greater than 0.5 not being respected when using the NumericRasterizer transformer.With tolerances like 0 and 0.5, I can see a difference in output. However, a tolerance value of 1.0 has no change in the output. Output from NumericRasterizer with tolerance set to 0.0, with the input vector overlayed  Output from NumericRasterizer with tolerance set to 0.5, with the input vector overlayed 

How can one rasterize a vector where all cells touched by a polygon/vector get drawn during rasterization?I have tried the NumericRasterizer, but I believe there’s an issue with tolerance values greater than 0.5 not being respected when using the NumericRasterizer transformer.With tolerances like 0 and 0.5, I can see a difference in output. However, a tolerance value of 1.0 has no change in the output (see below). Output from NumericRasterizer with tolerance set to 0.0, with the input vector overlayed  Output from NumericRasterizer with tolerance set to 0.5, with the input vector overlayed

I have opened an old workspace and all of the transformers in that workspace look like the first AttributeCreate in my screenshot. The Input port is hidden.When I copy the AttributeCreator, it looks different. The Input port becomes visible. It looks like the second AttributeCreator in my screenshot. When I add new transformers, they all have a visible Input port. When I open a completely new workspace and add transformers, they also have a visible Input port.How can I hide the Input port like on the first AttributeCreator in my screenshot?It seems like this cannot be achieved for any newly created transformers and workspaces. It only looks like this in this one old workspace of mine.There is an option under View that makes the Input ports visible on all of my Transformers. But it’s impossible to reverse that action. The option to make the Input ports visible disappears, once they are visible on all trans

I need to connect to a PostgreSQL database via an SSH connection. How can I do that ?Thank you

I had a hard time with the integration of GDAL and FME raster transformers. So, for the record, I posted a short tutorial:https://blog.fiorino.fr/index.php/2022/05/11/fme-python-api-working-with-rasters-and-gdal/

I need to update environmental data in our system. The input data comes from an ESRI Shapefile, and I am using FME 2021.2. The output is stored in Oracle Spatial.However, the system that reads the data from Oracle only accepts simple polygons. When I try to insert multipolygons, part of the geometry is ignored by that system.Is there a way in FME to convert multipolygons into single polygons while maintaining a 1:1 relationship between input and output features?If I use the Deaggregator, the number of records grows exponentially, which is not viable for our database.I also started working with FME recently, so I may be missing a simpler or recommended approach.What would be the recommended way in FME to transform multipolygons into valid simple polygons without increasing the number of features?

My long time transformer friend FMEFlowLogFileRetriever is not avaliable in FME 2025. The error asked me to use HTTPCaller to retrieve the job log file. So I used this rest point in the HTTPCaller’s Request URL with GET - https://my-fme-flow-hostname/fmerest/v3/transformations/jobs/@Value(jobid)/log. I tried the Basic and Web Connection authentications, but both gives the error “HTTPCaller (HTTPFactory): An error occurred while accessing the URL 'https://my-fme-flow-hostname/fmerest/v3/transformations/jobs/98/log'”. I have no problem publishing using the same Web Connection for the authentication, so don’t understand what the issue is. The jobid = 98 was run successfully on the FME Flow.

Hi,In a workbench, I created a Custom Transformer to loop batch of features.It is a workaround to prevent machine reboot on large feature datasets processing. At least, writing-along features in the Output to .csv. file help to recover features processed in PythonCaller and it is easier to resume for the remaining features… Here is the workbench, with Inputs:batch = 10counter = 54000 (total of features)loop = 0 (to keep track of iteration)offset = 0The Tester_counter transformer filter out batches 10 features at each iteration, sent to PythonCaller. At the end of processing the Tester_Loop tests if there is remaining features to process. Anything processed is sent to the Output and written in a .csv Writer transformer in the ‘Main’ workbench, whereas ‘failed’ remaining features from Tester_counter have their attributes incremented for variables loop and offset for the next iteration.However, when running the workbench, t

Hi there,We recently installed ArcPro 3.3 with the data interoperability extension FME version 2024.0.1.0.I have succesfully used WorkspaceRunner in the previous version of ArcGIS PRO 2.8.But now, when I use the WorkspaceRunner transformer to run a simple workspace the transformer fails to read the published parameter, it says "Specified workspace has no parameters or could not be read". The workspace I am loading into workspacerunner does have a published user parameter. Does this happen because of the new version of ArcGIS? Anyone else have this experien

I’m using the Esri Reprojector transformer with a feature reader that reads feature classes from an ArcSDE connection. They all share the same coordinate system and I need to transform them. I can do this successfully using three separate transformers but seems a bit silly. I’m assuming there is a way that I only need to use one and feed the three feature classes through it? I’ve tried doing this and adding a FeatureTypeFilter transformer but this returns no values if I click ‘Update’.  

Hi all,I’m currently working with several Revit files hosted on ACC, and I have put together a workflow that can read a specific ACC folder and identify a federated model. The process successfully reads and downloads the host file; however, I have been unable to download the associated linked files.Is there a way to use the AutodeskDocsConnector to access and retrieve the linked files as well?Thanks :)

Recently our network has been segmented which causes an issue with sending mail (emailer transformer) with FME Form from server X using an smtp mailserver which is in another part of the network. A firewall has been added in between these two. The firewall blocks mails being sent because fme form (20241202 - Build 24801 - WIN64)) supposedly uses an outdated version of OpenSMTP giving this issue: CVE Record: CVE-2020-7247Does anyone recognize this? It seems to be an old issue dating back to 2020. I would expect this to be solved in a 2024 version. I can't find anything in the documentation. Help would be appreciated since I know very little about firewall, mailservers etc. 

Hi,I am currently working on an assignment in FME and have run into some difficulties regarding how to structure the workflow to meet all the required criteria.In particular, I am unsure how to best combine the distance requirements. For example, how to approach buffering and overlay operations to ensure the area is within the correct distance from both the sandy beach and a major road, while also being sufficiently far from buildings. I am also uncertain about the most logical order of steps to avoid creating an unnecessarily complex workflow.I would really appreciate any advice on how to approach the structure of this task in FME in a clear and efficient way. I am mainly looking to better understand if there is any other transformers who would make the workflow more efficient?Thank you in advance for your help.Best Regards Anton

I am using the Precisely MapInfo Extended TAB Writer to generate a MapInfo table that is over 2GB (total) in file size. I estimate it should be between 3GB and 4GB in size.I am indexing five attribute columns in the table.And the process of indexing the attribute columns is extremely slow.When I do not index any columns, it takes about 20 minutes to Write the tableWith indexing the columns, it can run for twelve hours before I halt the process in frustration to check my process.Does anyone have any experience of this? Is the Precisely MapInfo Extended TAB Writer just bad at generating indexes?