Authoring

Hi.  I’ve created a workbench that reads information from a PDF file, processes that content, then writes it to a database.  This works as expected, either when run locally or from within FME Flow.  However, it only processes one file at a time.  So, to rectify this as a local process, I have created a separate workbench that uses a ‘Directory and File Pathnames’ reader to read PDF files from a named folder, then passes these individually to the original PDF-processing workbench (via a WorkspaceRunner).  This, too, works as expected.  However, being local, it’s slow.So, it would be good to replicate this, so that it instead makes its requests to FME Flow, via an FMEFlowJobSubmitter.  And...this is where I get stuck, as it requires each PDF file to be uploaded to that server, in sequence.  Looking at the documentation, it seems that an HTTPCaller is the preferred way to achieve this, but I’m not sure that I understand how this needs to work in practice.  So, has anyone done anything sim

I have been using FME2026.2 -26333 - for all my current projects which includes analysing larger datasets (2-5mil records).  These aren't really that large.I have found when building the analysis with data cached FME can be highly unresponsive and lag when trying to modify the existing transformation.  Adding a new transformer, modify an existing transformer, even moving objects on the canvas.  The system will lag and can often crash.My laptop is the highest spec our company supports - Dell Precision 7680, 1Tb HD, 6gb GPU, 32gb RAM.  Even clicking too quickly after the transformation is run for that transformer can crash FME.   Surely I’m not the only person with this issue?  Stability in this version seems unbelievably poor which makes me wonder if this is known or overcome somehow.

 I have some strange behaviour that Im hoping Safe can confirm, I have built a very simple loop in 2026.2 using the new looping functionality, it runs fine every time.  I then encapsulate that into a custom transformer, thats where the weirdness starts.  Running with caching OFF everything behaves itself...I then turn caching ON and the loop on the canvas behaves as expected, but the custom transformer containing the SAME loop doesn’t…Inside the custom transformer…The Loop doesnt initiate, when caching is ON, I cant work out why?I have attached the workspace.

Hi. We're looking for some clarification on the virtual memory values being reported by FME Form and whether they are expected. On a Azure Virtual Desktop (AVD), Windows System Information shows:Disk: 128.00 GB Drive C: Total: 126.45 GB Drive C: Free: 36.37 GB Physical RAM: 64 GB Page File: 9.5 GB Virtual Memory: 73.5 GBHowever, FME Form 2026.2.1 reports the following in the logs:System Status: 36.38 GB of disk space available in the FME temporary folder (C:\Users\LOCAL_…...\Temp) System Status: 63.99 GB of physical memory available System Status: 255.98 GB of virtual memory availableThe physical memory figure appears reasonable, but the virtual memory figure is considerably higher than the virtual memory reported by Windows.Our concern is whether this is simply a difference in terminology or metrics, or whether FME is using a different definition of "virtual memory" than Windows System Information. If FME genuinely believes more memory is available than can actually be committed by th

I’m using StringPairReplacer to replace some tags (inside $ chars) in a snippet code to their default value, which fails.Since SPR doesn’t accept (dynamic) variables for replacement string, only parameters, I’ve enclosed it in a custom transformer, and using its parameters to transfer the value of the replacement string.But it still fails with“@Tcl2 -- failed to evaluate expression `Replace_replacement_tags_StringPairReplacer_2_fb7401f4_4759_4fdf_9add_13b9915a14686_replace' -- char map list unbalanced”The replacement value is given below in blue, and it works if I insert it as a fixed string.$schema$ dbo $table_name$ table_name $column_name$ column_name $data_type$ VARCHAR(20) $null$ NULL $collate$ Latin1_General_CI_AS $default$ NULLAm I fighting some TCL or FME escaping, or could it be a bug in SPR ?Fails in both 2020 and 2026.Cheers.

Hello,I recently noticed some unexpected behavior with the Feature Merger in FME 2026.2.1.I have a workspace where a Feature Merger combines data from three different streams. Under normal circumstances, everything works as expected. After successfully testing the workspace locally, I published it to FME Flow.Today I noticed that the workspace failed during execution. When I opened the workspace in FME Form and ran it with Feature Caching enabled, I was surprised to see that the Feature Merger did not merge features that clearly shared the same key values:What makes this even stranger is that if I right click the transformer and select "Run Selected Objects and Any Required Upstream Objects" a second time, the merge completes successfully without any changes to the workspace or source data.Has anyone experienced similar behavior with the Feature Merger in FME 2026.2.1? Could this be a caching issue, an optimization problem, or possibly a defect?Thanks in advance for any insights.

Hi,With the new SQL Server Reader is it possible to simulate the older non spatial reader?I have 3 million records with line geometry.  I only want to read the attribute information and not the geometry. I can’t spot how to do this in the 2025.2 Reader.  Is it not possible or have I missed something simple?Thanks, Annette

I have just migrated my FME Flow installation to 2026.1.1.  I have ArcGIS Pro installed on the server.  I am having a strange issue with workbenches failing due to licensing.  I have multiple workbenches which use ESRI geodatabases and are scheduled to run repeatedly 24/7.  At least one of them fails intermittently due to:Unable to perform any licensing related tasks. Please check your Esri ArcGIS Compatibility workspace parameter, found in Navigator -> Workspace Parameters -> Spatial. This 64-bit version of FME requires ArcGIS Desktop along with 64-bit Background Geoprocessing; ArcGIS Pro; or ArcGIS Enterprise. Refer to this this article http://fme.ly/qua for more detailsThe compatibility parameter in the workbench is set to ArcGIS Server/ArcGIS Pro. It will run successfully 3 times, then fail twice. No pattern to the failures.  How can a licensing issue be intermittent?  I have multiple other workbenches using esri licenses that run successfully.Has anyone seen this before?

I quickly became a big fan of the recently introduced in-canvas looping functionality in FME. Here is one workflow where loops make things particularly elegant: building geometry from list attributes without first exploding the list into separate features.In this example, each list element contains a coordinate pair such as [500,611]. Before the feature enters the loop, ListElementCounter records the number of elements. On each iteration, we use the iteration number to retrieve the corresponding pair, extract its X and Y values, and create a vertex.The expression used for the JSON Document parameter is:@Value(_list{@Evaluate(@Value(_iteration_num)-1)})We subtract 1 because FME list indexes start at zero, while the loop iteration number starts at one.Every generated vertex leaves the loop and goes to LineBuilder. At the same time, the feature goes to a Tester, which compares _iteration_num with _element_count. While the values are different, the Failed port sends the feature back for an

I have a QGIS .model3 file that uses PostgreSQL tables as its inputs and writes to other PostgreSQL tables as its outputs. I am having trouble getting this model to run as part of an FME Workbench.What I have done so far:Copied the tool parameters from the QGIS graphical interface (Advanced > Copy as qgis_process command) Edited the command so that “--” only appears in front of the first parameter Added “” around each PostgreSQL parameter as they contain spaces and special characters and were causing Invalid parameter errors in Windows command prompt Thanks to these changes, running the tool with qgis_process in Windows command prompt no longer throws invalid parameter errorsThe tool fails silently through the SystemCaller transformer in FME.In Windows command prompt, the parameters now seem to be read properly but I am now getting “File could not be found” for PostgreSQL addresses.Still in command prompt, dialing psql first with my database credentials makes it say \PROGRA~1 is an

I have ArcGIS Pro Data Interoperability Extension.  Within the last couple of days, it keeps stalling at the start screen “Checking License”.  I know I have a license!  And it works sometimes…without any trouble.  If I Restart my PC, it opens fine, but if I try to run a reader (excel, csv), that is stalling as well.  Even though these are small files, they just don’t come in.  I’ve tried starting/stopping/pausing/purging, etc.  The only thing I’ve been able to do to get the Extension to open again, is to do a cold boot of the laptop.Is there some other cache somewhere I need to clear out?  (I’ve cleared the cache from within FME Form itself) several times, but then end up rebooting anyway.  I’m relatively new to using FME, but these files are quite simple with only a couple thousand records.  I tried to read both as .xlsx and .csv today.  Could this be an issue with a log file filling up?  Thanks for any tips/suggestions.

I need to download multiple files and folders from SFTP to a SharePoint folder. The SFTP connection uses a custom port and a private key file (.ppk) instead of password. How to achieve this via FME 2026.2?

Hi all.I just created a new custom transformer, and need to apply it multiple times (10-20 times).It basically just has a name and three parameters, all of which can be deduced from a database query.So I thought it to be a good idea to create and paste the definitions from a query result (using some sql magic). Instead of copy/pasting the many transformers as I would normally, editing each by hand.So I experimented at bit first.I can copy a transformer from FME Form, and paste it right back. I can also paste the definition into a text editor, as the copied content is just text (FME file syntax).But when I just copy the complete transformer (without changes) back from the text editor (Ctrl-A,Ctrl-C), FME Form will not paste it into the workspace.I used Ditto clipboard to look for differencies, and I can barely find any. The two attached images shows the properties for the original and the text editor copy.original from FME Formtext editor copyThe only difference I can find are the sligh

One of my clients is using FME on Linux with only the command line interface and is having some trouble importing database connections. Specifically: they don’t know how to do it on the command line.I can’t find much in the way of documentation for this. I suspect that if it is at all possible it may be through the APPLY_SETTINGS command, but other than that I can see that that exists, I can’t find any indication anywhere on what those settings may be and how to set them.Any clues? Anybody with an old manual somewhere? I do have old installation cd’s and dvd’s going back to FME 2005, but I don’t think I have a computer with a cd or dvd drive anymore 😅 Hoping one of the other long-time users has something stashed away...

Hi,I am having an issue setting multipart uploads with httpcaller in 2026.2 FME Form.Please see embedded video, but basically the gui is glitching and I cannot see my values for the multipart strings correctly. 

Hi everyone,I’m publishing hosted layers to ArcGIS Online (AGOL) using the Esri ArcGIS Feature Service writer. My issue is that the map extent for the item in AGOL is at 0,0 (aka middle of the ocean). I’m not sure what I’m doing wrong to cause this problem. I have attached screenshots below of my workflow and writer settings. I reproject the layer to ESPG:3857 as recommended. I have also tried with ESPG:102100 and no luck. Workflow to publish to AGOL Write parameters in Workbench 2025.2The current map extent in AGOL:My goal map extent that is zoomed into the layer: 

If I have an MS SQL Writer that I have disabled for whatever reason, all the features in the workspace turn grey. Perfect, is a very good indication that it is disabled. However: if I import another table, that writer is added to the Main canvas and to the Navigator in full color!Which caused me quite a bit of head scratching this morning: why was nothing written to the database? It seems to me that this is a bug in FME Form (FME(R) 2026.1.2.0 (20260512 - Build 26124 - WIN64), but maybe I am overlooking/overthinking something…To reproduce: Add an MS SQL Server Writer, use Import from Dataset to select a table. Right-click on the Writer in the Navigator, and select Disable [writer_name]. Table on the canvas turns grey, as it should. From the menu select Build - Writers - Import Feature Types. Select another table, click OK. Now I see one grey table and one in full color… 

Hi, my name is Leen.I have connected several items from an Excel-column in FME and gave them consecutive itemId’s and placed them in a XML-file. this is necessary to import these itemId’s in Teamcenter.Now I want to connect other items from another column from the same Excel-file into my XML-file with the use of the follow-up consecutive numbers that have been created earlier.

I often find myself reusing subprocesses from my main workspace, either for debugging or for creating similar workflows with small changes in a new Workspace. I'm aware that FME has a ton of useful shortcuts, so I was wondering if there's one that lets you automatically select all the necessary(only!) transformers that belong to a previously used process. The idea would be to quickly copy and paste the used chain into a new workspace without having to manually select every necessary transformer or just copy paste all. Is there a shortcut or built-in feature for this, or is manually selecting the chain still the best approach? I feel like this is most likely highly specific to my own taste but I wanted to ask nevertheless since the main process got way too big :DThanks in advance! :)

Hello 😊I recently updated FME Form to version 2026.2 and started experiencing an issue when connecting to an Enterprise Geodatabase using an SDE connection file configured for Operating System Authentication.FME returns the following error:"Failed to connect to Enterprise Geodatabase using connection file 'path_to_the_sde_file' due to missing credentials. A connection requires either Windows Authentication or both a user name and a password."The SDE connection itself appears to be valid. If I open the same .sde-file in ArcGIS Pro, it connects successfully and shows the correct authentication type in the connection properties.This workflow worked without any issues in previous FME versions and only started failing after upgrading to FME Form 2026.2.Has anyone encountered this problem before? Is this a known issue in 2026.2, or is there a workaround available?Any help would be greatly appreciated.

Hi everyone,After upgrading our FME Flow instance from 2025.2 to 2026.2, a workspace parameter with dynamic values stopped working in a Workspace App.When opening the app, FME Flow shows the following error when testing the dynamic parameters:Loading from WorkspaceKartenautomatisierungen/Kartenautomatisierung_ZUBA_PreScript_2.fmwError loading schemaUnable to generate dynamic parameters from source file.Error was:E:\appdata\fme\systemshare\resources\system\null\null\FeatureWriter\ZUBA.jsonPlease ensure the file exists and is readable. A few observations:The same setup worked without issues in FME Flow 2025.2. The workspace itself executes successfully. It loads the JSON after the workspace is done.  The error only occurs while loading the dynamic workspace parameter. The path contains system\null\null\FeatureWriter\ZUBA.json, and the corresponding folders do not exist. I never configured FME to write schema files to that location.Has anyone seen a similar issue after upgrading to 2026.2

Quick question.Are there alternatives to using AttributeManager to manually sort attributes by name?Sort of an AttributeSorterByName or AttributeSortedReexposer transformers ?Cheers.

Hi all.I’m experimenting with calling WFS services with custom filtering expressions, including spatial filtering.According to the filtering documentation, spatial filtering is possible with GML, but it doesn’t seem to work in FME ?The geometry column is present but has no value in the output from FeatureReader, and FeatureReader has its own spatial filtering parameter separate from the “normal” filtering parameter (using <Filter/>)Am I correct in assuming, that FME effectively prohibits any spatial filtering via the “normal” filtering parameter, and requires you to always use the special FeatureReader spatial filtering parameter ?And maybe this is neccesitated because FeatureReader removes the WFS geometry column values ?Thanks for any insights.

Hi.This may be a simple or a complex question, I’m not sure which.I’m attempting to use a dynamic expression, that I construct dynamically from values read from a file or table, and evaluate it on all relevant feature read from a database or file. Including AND/OR etc.How can this be achieved in Workbench ?I’ve tried to utilize ExpressionEvaluator with the expression @Value(@Value(expr)), but it doesn’t return any value in the “_result” variable.What am I doing wrong, and is it even feasable ?Are there better ways to achieve this ?Using 2025.2.4Cheers

I am looking at this Article for setting up Azure OAuth Connection for Data Bricks, the article is for Reader, but I assume the same set up will work for the Writer.   We have an Azure app already set up for the FME SharePoint Reader and Connector, can the same Azure app work for both SharePoint and Data Bricks ? are there any best practices about using the same Azure app for both or making separate Azure Apps for each ?   Any suggestions or cautions, things to watch out for, etc,  appreciated  https://support.safe.com/hc/en-us/articles/27147444106381-Setting-up-a-Databricks-Reader-Azure-OAuth-Connection