Best Answers

Just upgraded from 2022 to the latest version for Flow and Form. Whenever I publish a workbench the publication works then shortly after the log gets the message “The web connection 'FME Account' could not be loaded”We don’t have a Web connection “FME Account “ and the task has completed anyway. Any ideas where this is coming from?  

There are situations when I want a variable or parameter to determine whether a certain process is run or not. I could add the variable as an attribute to each feature and then test for the value of the variable. But that way the variable controls the process for each feature, when I already know that I don't want put any of the features through the process. And it feels like creating an unneccesary amount of extra data (even though temporarily), which I have to remove later on.I've come across this before, and now I have a case where I want to run part of my workspace once a week and another part every day. (An alternative is to distribute the processes across separate workspaces, of course.)What I'm thinking of is functionality like the feature reader has: it only reads features when triggered by an initiator. But I can't think of a way to do this with other transformers.Does anyone know of a way to achieve this in FME Workbench / Form?

Hi,I recently upgraded from 2025.1.1.0 b25615 to 2025.1.3.2 b25646 and noticed that whenever I open a .fmw file, a new folder, “.sentry-fme” is created in the same directory.When I started FME from my desktop shortcut, a “.sentry-fme” folder was created on my desktop once the program was up and running.Could we get this fixed? I’d rather not bloat my directories with temporary files if I can help it. I’ve since downgraded back to 2025.1.1.0.Thanks,

Good evening,I am wondering what the community thinks about this situation:FME Form 2024.1.3.0 (20240927 - Build 24627)Writing to Esri Enterprise Geodatabase (SDE) 10.9.1 - MS SQL Server 2022I have almost 700,000 records that need to be written out to our SDE into 1 feature class.Currently I have 1 Writer using the Esri Geodatabase (ArcSDE Geodb) format.The 1st attempt to write the data failed after 22 hours because I had a loc on the destination feature dataset.  doh!On the 2nd attempt to write the data to the SDE, the job has been running for 12 hours.  I cancelled the job, and below are snippets from the logfile highlighting how many records have been written, and how much time passed.  This shows a pattern that: as time progresses, the processing time is becoming exponentially longer.  I am thinking of breaking the data into 100,000 chunks, with the first chunk first truncating the table and then loading, and the sequential chunks appending to the table.Is this performance hit expe

I am trying to validate some properties of objects in DWG files, and some of the things I’d like to check require me to know what paper space viewport an object is visible through, but I have been unable to get the autocad_viewport_handle attribute to show up.I’ve tried a few combinations of reader parameters, including checking all boxes in the “Model and Paper Space Options” section, unchecking some options while leaving “Read Model Space Entities Through Viewports” checked, and tried reading multiple DWGs, but so far I have not had any luck. I am thinking this must be a “me” problem, but I’m kind of at a loss. Is there a certain set of parameter settings required to get autocad_viewport_handle to show up? I am setting the reader to expose the attribute, but so far, it is always missing. I’m using FME 2025.0. 

 I am trying to make a custom top level resource folder work, but i can't write the output of my apps and workspaces to the folder.  If i try to write output to the custom folder, it fails almost immediately, and when i inspect the logfile it just says Log Error The log file for job ID "816" does not exist.Did i forget some steps in the configuration of the custom folder? Or what did i do wrong? Thanks in advance for the help!  For reference what i did:Adding Top-level resource folder in FME Flow | Community

Hi I am uploading data to S3 using the S3 connector. However, empty folders are not being uploaded, even though I have selected “Contents only” as No. I need to upload some projects that contain empty folders and subfolders as well.!--startfragment>  Many thanksNasir

Is FME able to read the MicroStation DGN file geographical coordinate and use this information when creating esri geodatabase features?  I see several tutorials and methods that seem to suggest CAD data such as DGN (MicroStation Design Files) do not contain geographic Coordinate systems but is not totally true.  Design files do or can have coordinate systems, and most do when relative to Transportation Civil Engineering Design. 

HiI have a a workflow exracting data from an API. I have got almsot everything I need from the JSON but need to get as single description field. The JSON structures breaks the description into sever parts all nested within the description part. The number of parts varies by feature.I essentially need all the text elements in the image below  When I use a JSON exposer it asks me to specificy each indexed element, but I do not know how many there will be in any given feature  I imagine there is something really obvious I am missing hereClosest I have got is adding another JSON extractor using the query:json["fields"]["description"]["content"][*] is giving me:{"type":"paragraph","content":[{"type":"text","text":"This Jira ticket is for all activities related to the implementation of 365 – "},{"type":"text","text":"Office.com","marks":[{"type":"link","attrs":{"href":"http://Office.com"}}]},{"type":"text","text":" rollout."}]}Or using:json["fields"]["description"]["content"][*]["content"][*

Hi FME users,I am creating an automatic workflow to download GTFS (Transport Data) by HTTP caller. The zip file contains few CSV files which I need to read them separately for my workspace. The issue is I cannot reach each file through the zip folder by using Feature Reader. I want to avoid unzipping manually as I want to create an automatic workspace. Can you please advise on how to solve this issue? Appreciate your help.  

Hi,I've been trying too use autocad_xref_path and autocad_xref_name to insert file (another dwg) into REALDWG writer feature type. It should be basic. But it's not working. No warnings, no errors.Did anyone succesfully managed to do that? Can you explain me what am I doing wrongRegards,Želimir

I’m running FME 2024 soon to be 2025.My workflow is reading tables from SQL Server and writing them to Snowflake.Can someone please advise if it is possible to have the workflow automatically handle new columns being created on the source table?I don’t want to have to publish a new version of a workflow just because a source table has a new column added to it.Thanks in advance.

Hi all,I’m trying to read a shapefile that is stored inside a ZIP file which itself is embedded inside another ZIP. I want FME to read it directly from the download URLI tried with https://url/data.zip[shp.zip] but did’nt work. any idea if its possible?Thanks

I'm trying to reproject a DWG file from AGD 1966 ISG 56 (EPSG:102077) to GDA2020 Zone 56 (EPSG:7856). I'm using the original DWG as a template to preserve layers, colours, line styles, etc.I've set up a workbench to handle the reprojection and all the data looks good and in the correct location when viewed at the final step in the inspector:However, when I open the output DWG, most of the data is correctly reprojected (matching what I see in the Inspector) but some features remain in the original coordinate system, thus showing up in Antarctica:Zoomed in at the portion in Antarctica:I've tried several troubleshooting steps:A simple read/write workbench Writing to DXF instead of DWG Using the RealDWG reader/writer Using default DWG reader settings (no customisation)I saw this question is similar to mine, but I didn’t have any luck writing to dxf instead:None of these things have fixed the issue. It seems to mostly affect text features, although some text is reprojected correctly, so I'm

My fme flow 2025 server setup requires proxy settings to allow access to other ressources in the net.This seems to cause trouble with the Elastic Writer (Elastic Cluster in this case).When switching the proxy off, it is no problem to write content into a elastic index.Switching the proxy on and putting the elastic server in the “no proxy for” list, leads to connection attempts to the proxy server and then timeouts, like here (10.150.0.77 is the proxy server):2025-10-28 20:45:44|   1.2|  0.0|INFORM|ELASTICSEARCH_CLUSTER Writer: Feature Type Parameter 'fme_match_id_attribute' has value '_id' for feature type 'adressen_4326'2025-10-28 20:45:44|   1.2|  0.0|INFORM|ELASTICSEARCH_CLUSTER writer: Opening cluster 'http://geogdiesp001.xxx:9200/'2025-10-28 20:46:05|   1.2|  0.0|ERROR |ELASTICSEARCH_CLUSTER writer: JSON parsing error: boolean expected (at offset 2)2025-10-28 20:46:05|   1.2|  0.0|ERROR |ELASTICSEARCH_CLUSTER writer: Failed to connect to 10.158.0.77 port 80 after 21004 ms: Could n

I have been trying to find a good use for the ‘ai assist’ function. I don’t think it has ever succeeded in providing a useful answer. Have any of you had any luck?Or do you also copy paste to a real ai?In this example I needed a SQL Query edited, but my request was invalid for no reason. When I post in any other LLM I get a thorough response and description of this and that.My overall questions is probably, why is the FME implementation so poor? Is this spelling the early signs of the end for FME? 

Hi Everyone,I’m new to FME. Currently, I use FME to convert IFC4 LOD 300 to CityGML 2.0 LOD 4. It happened that a long railing can make the size of one CityGML feature exceed 1 GB, and it’s challenging to import it into 3DCityDB. I have to divide the railing into pieces so that each CityGML feature won’t exceed 200 MB. As shown, the railing has 1270 parts of MultiSurface. I’m thinking of splitting it into 1270 CityGML objects.First, I tried the “Deaggregator” transformer. But when it is written to CityGML with other feature types, the divided geometry cannot be viewed. The file size is small, so I believe the divided geometry wasn’t written into the CityGML file. However, if I only convert the divided geometry into the CityGML file. The file size is 1.4GB, and it can be viewed in FME Data Inspector. Here is the screen capture of my workbench.I have no idea what was wrong. I was thinking it could be due to the geometry structure being changed. Originally, it was Geometry->MultiSurfac

I try to create tables where schema definition is from SQL Server into PostGIS. Schema onlyFor this i have a simple workflow. But for the life of me i do not understand why my tables are not created despite FME report everything went well?String replacer is just changing the name of fme_freature_type_name to remove the standard “dbo.” table qualifyier from SQL Server.   Writer setup Log                           Features Written Summary=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-rplanforslag_alternativ_forslag.RpAngittHensynGrense                         1rplanforslag_alternativ_forslag.RpAngittHensynSone                           1rplanforslag_alternativ_forslag.RpArealformålOmråde                        1==============================================================================Total Features Written                                                       3=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=--~-~-~-~-~-~-~-~

I feel like everytime I try the PythonCaller I really struggle with the scripting and usually switch to a start up script to avoid it. Can anyone help me figure out why I don’t have any outputs when I run this in my PythonCaller.  import fmeimport fmeobjectsimport reclass FeatureProcessor:    def __init__(self):        pass    def input(self, feature):        """        Processes a feature to extract and set address components.        Args:            feature (fmeobjects.FMEFeature): The feature containing the address attribute.        """        # Retrieve the address attribute from the feature        address = feature.getAttribute('ADDRESS')        # If the address is missing, set all related attributes to empty strings        if not address:            for attr in ['Suite', 'AddressNumber', 'StreetName', 'StreetType', 'Direction', 'Street']:                feature.setAttribute(attr, '')            return        # Split the address into suite and core address parts        parts = [p.

Is it possible to use FME Flow Hosted with a Remote engine located on an internal network. Thinking that this configuration would minimise FME Flow core maintenance, Simplify firewall rules, provide access to internal network resources like file system shares and databases... If it is possible, is there any documentation on how to setup this configuration.Cheers,e. 

Our current FME Flow configuration uses 4 standard engines and 2 queues. One queue is set to high priority and is dedicated for workspace apps for quick turnaround jobs, and the other is the default queue with medium priority. One engine is dedicated to the quick turnaround queue, and the other 3 are dedicated to the default queue. We have a single job routing rule that assigns jobs that are sourced from workspace apps to the quick turnaround queue. The question is, if the default queue has no jobs at the current time and there is already a job running in the single engine dedicated to the quick turnaround queue, can a workspace app job be moved to the default queue, instead of waiting for the current job to finish? I’m guessing this is a simple ask, but I’m missing how to accomplish it.

Hello FME Community, Salesforce recently announced the retirement of their SOAP API login().  Are there plans to replace the current Salesforce readers and writers or update them to support Salesforce’s OAuth 2.0 Flow?  Currently the readers and writers use /services/Soap/u/51.0 which will be retired in the Summer 27 release.  I’ve used these a good bit in production workspaces, so need to start mapping out some transition plans.  For retrieving data, the SalesforceConnector will work, but it’s nice to be able to write and update data with a drop in transformer rather than setting up HTTP callers with Salesforce’s REST API. Any thoughts on the current plan for the Salesforce readers/writers or ideas on how to make this transition seamlessly? 

Would be great to be about to output Dynamic blocks to AutoCAD, things likeVisible states and Flip setting would be great, Distances for stretch, rotation and Alignment parameters might be a bit harder to achieve.

Hi FME Community,I came across the following notice in the AWS S3 API documentation about the Owner object:“Beginning November 21, 2025, Amazon S3 will stop returning DisplayName. Update your applications to use canonical IDs (unique identifier for AWS accounts), AWS account IDs (12-digit identifiers), or IAM ARNs (full resource naming) as a direct replacement for DisplayName. Between July 15 and November 21, 2025, you will begin to see an increasing rate of missing DisplayName values in the Owner object.”Reference – AWS DocumentationCould you please confirm whether this upcoming change will (or might) affect how FME Flow interacts with Amazon S3 — for example, any components that depend on the Owner object or its DisplayName field?Environment: OS: Ubuntu FME Flow 2023.2.1 and 2025.1.1 Thank you in advance for your guidance!

I am trying to read in non-spatial data from Microsoft SQL Server. I need to read in three Tables and three derived Views. I have set up a connection using the Microsoft SQL Server Spatial (JDBC) format, and I have the connection to the main folder in the database working fine. In the Parameters > Constraints, I am only offered the option of ‘Tables’, and this does allow me to read in the three tables in the database perfectly. However I cannot find a way to access the Views which are in a different folder in SQL. Is Reader limited to just reading Tables? The FME documentation says “The Microsoft SQL Server Spatial (JDBC) reader and writer modules provide FME with access to spatial and attribute data held in Microsoft SQL Server database tables.” - with ‘tables'; being specifically mentioned, rather than all folders within a database.Is there another way of reading in Views? Or do I need to go back to the database creator and get him to do something else.