Shape the future of FME with your ideas
New and Open ideas are awaiting review or have been reviewed by our Product Management team, and are open for commenting and voting.
Asking for an improvement to FME Flow Apps, to better support ArcGIS Enterprise or ArcGIS Online content. Currently web mapping tile services are supported, while these are useful for basic public mapping scenarios, it can be limiting for enterprise users, like ESRI customers who rely on managed GIS content, authoritative operational layers, and internal esri basemap standards. Workflow for Geometry Picker….In FME Form under user parameters there is an input called “Spatial Geometry” type.Define & Select spatial Geometry user parameterWhen this parameter type is published to FME Flow and a ‘Workspace App’ is created, under the ‘Parameter Defaults’ there is an option for Map Configuration, as outlined in this helpful article Using the Geometry Parameter – FME Support Center.proposed UI improvement to Map Configuration in Flow apps - add ArcGIS map services Why add ESRI map support to Flow Apps?Many organisations already manage authoritative spatial context through Esri platforms. When a Flow App requires a user to draw or select an area of interest, the user often needs to see operational reference layers such as boundaries, tenements, assets, work areas, inspection zones, environmental constraints, or approved operational footprints.Without access to these layers in the Geometry Picker, users may need to manually cross-reference another GIS application, which can reduce confidence, increase error risk, and make the Flow App feel disconnected from established enterprise GIS workflows. Desired spatial geometry parameter functionality:Support existing Esri webservice/connections already configured and authenticated in Flow to come into the Flow spatial geometry parameter Allow Flow App authors to configure Esri REST 1. map and 2. feature services as selectable or reference layers in the Geometry Picker. Respect secured services where authentication is already configured in FME Flow, including (arcgis) web connections or stored/encrypted web service credentialsThe old htmlreportgenerator supports the consumption of “feature layer URLs”, even multiple layers if they are comma separated. HTML report generator support Esri leaflet - feature service URLs shared openly I posted on this in the past, and the community wanted the transformer to see more support added as well. HTMLreportgenerator ESRI leaflet - feature layer URL more info needed | CommunityOne issue I anticipate might be challenging 😥, is that the latest ARCGISFEATURESERVICE safe.esri.agol package only supports content that is my content or shared to groups. The above functionality all assumes the esri REST map/feature services are shared to groups. But through the use of web connections to arcgis content, supporting group shared or originisational shared content like basemap services could be explored?the latest ArcGIS feature service reader > its parameters to item sourceThe Background Maps available in FME Form, would be very complimentary if added to FME Flow and Workspace Apps!Much of the functionality could be used in a similar manner 😎in FME form we have ‘background map’ support… how about this in Flow apps? You might find other ideas similar to this onehttps://community.safe.com/ideas/geometry-picker-basemap-using-arcgis-webmap-29298 https://community.safe.com/ideas/set-a-default-map-for-geometry-picker-in-fme-flow-app-29958 https://community.safe.com/ideas/flow-apps-customize-geometry-parameter-search-37946
I am revising/ simplifying the idea it would be nice if to start with the path reader would allow exclusion of write protected (received an error and failed), hidden, system, invalid folders/files. secondly I currently use a tester to exclude many folders that I am not interested in scanning. But this requires a read to get the features through the tester. It would be optimal to tell the path reader to use a list or sheet of folder names to exclude from the read. I know this sounds perf expensive but when scanning several hundred thousand folders the payoff is NOT to read them all before exclusion. leaving this confusing original section below:I have a question in community forum for a flow issue that cropped up. I previously had a ticket with my own IT with an issue that falls into this same category and now a third iteration.in each instance the reader fails and does not continue. whether in form or flow.instance:we receive a zip file that has been incorrectly named. When it is unzipped the folder name ends with a period. (eg. my folder-0name.it doesn’t\matter if there is a\period in the folder it only matters if the folder name ends with period.) this causes the reader to fail and cancel running in both flow and form. I have not identified why the second instance occurs but it is odd that if i am using a reader and it fails with ‘Media is write protected’. see this question here. I understand the 3rd instance is an error on our network. but still can’t the reader detect that it can’t really (with only symlink references) read this and just circumvent the folder? (the error is - PATH Reader: Error reading path '\\Tx-mio-mis01\fw-shared2\SURVEY\CHus\HM_PKG_9_LOWLINES\0_GIS\FOOTAGE_TABLE.xlsx' with error message: 'directory_entry::status: The symbolic link cannot be followed because its type is disabled. Response from my IT dept… From: Helpdesk <helpdesk@topo.sur>Date: Wed, 19 Aug 2026 16:13:56 (UTC -05:00) These are not corrupted. They are a new form of archiving that we have begun implementing. Are they giving you troubles? create specific exclusions. for instance maybe with regex. (e.g. _ARCHIVE) do not read
Being able to reorder the attributes on a transformer is awesome. However, it's also tedious. Repeated right-clicks and selecting "Move Up" or "Move Down" in a long attribute list is painful.It would be great if you could drag/drop attributes to change the order. So instead of clicking and dragging anywhere on the transformer to move it on the canvas, clicking and dragging an attribute should allow it to be moved up or down the list. Then transformers would only be moved on the canvas if you clicked the top of the transformer (where the name is displayed) and dragged it around.
See https://knowledge.safe.com/questions/68957/get-the-last-index-of-an-array.htmlI now know how to access the last element in a list in a single statement thanks to @david_r's answer to this question. So I no longer have to count the list elements then subtract 1 from it before getting the value of the last element in the list. However, in the process of finding out how to do this, I discovered that the ListIndexer allows you to specify a negative number to start from the end of the list.I haven't yet had a scenario where I needed to get (for example) the 3rd element from the end of the list but these situations are bound to arise at some point! So I'd like to suggest that negative numbers be allowed e.g. _list{-3} to start from the end of the list rather than the start.
Request: Please add parameter options for for the Esri Geodatabase Open API reader that allows us to ignore the FeatureDataset path to featureclasses I have an esri GDB with featuredatasets (see featuredataset image)With FME using Esri Geodatabase (file geodb) reader and an ESRI license... it nicely drops the featuredatasets for feature types.... and keeps the featureclass nameDistributionCarrier_HV_OHUse the Esri geodatabase (Open API) reader however and you get ... featuretypes with the featuredatasets in the path. No parameter option to ignore themxxxxxx_Distribution_Carrier/DistributionCarrier_HV_OHI workaround for the OpenAPI is to use stringreplacer to format them such that I can get to the featureclass names, Thanks to AI assist in 2023.1, I can now write regex !
Hi Safe Software team,We are looking to improve our automated data pipelines by integrating alternative object storage like Cloudflare R2 and Hugging Face with FME but the current Amazon S3 Compatible Storage connection is too strictly tied to AWS defaults.Could you expose a few standard SDK overrides in the S3Connector and Web Connection configurations? Specifically, we need: Free-text Region selection (e.g., to allow auto for Cloudflare R2). Addressing Style toggle (Path-style vs. Virtual-Hosted). Support for namespaced Endpoint URLs (to prevent FME from truncating trailing paths like /my-org for Hugging Face). Advanced Checksum controls (ability to disable strict AWS payload checksums). Adding these standard parameters would seamlessly unlock zero-egress and AI workflows directly within FME Form and Flow. Thanks :)
As of FME 2026.2, the new IFC_API Reader converts IFC geometry of type IfcFixedReferenceSweptAreaSolid into BRepSolids within FME (see IFC-to-FME Geometry Mapping). When these geometries are subsequently written using the IFC_API Writer, they are therefore written as IfcFacetedBrep.I would like to suggest native support for reading, preserving and writing IfcFixedReferenceSweptAreaSolid.The main issue with converting this type of geometry into a BRep is that a relatively compact parametric representation can become a very large explicit boundary representation. A swept-area solid can essentially be described using a cross-section, a directrix and an orientation rule, whereas a BRep needs to explicitly store the resulting vertices, loops and faces.As a practical example, FME currently does not appear to provide a native geometry type for sweeping an arbitrary profile along a path. For a railway use case, I therefore created the rail geometry manually in a PythonCaller by placing the individual rail-profile points along the input line and connecting them to form the final solid. While this works geometrically, the resulting object has to be represented inside FME as a BRepSolid and is therefore written to IFC as IfcFacetedBrep. This can result in considerably larger IFC files compared with representing the same geometry as an IfcFixedReferenceSweptAreaSolid.In one test containing only around 200 metres of rail geometry, the FME-generated IFC using IfcFacetedBrep was approximately 23 MB. After reconstructing the same rail geometry as IfcFixedReferenceSweptAreaSolid, the resulting IFC was only around 150 KB. While the converted file was created as a proof of concept through post-processing, the difference clearly illustrates the potential impact of the geometry representation on file size.Ideally, FME could provide a native swept-profile geometry type that retains the relationship between the profile and the path, instead of immediately converting the result into an explicit BRep. A comparable concept already exists for IfcSweptDiskSolid, which can be represented in FME using the Pipe geometry type.Such a geometry type could then allow the IFC_API Reader and Writer to preserve or create IfcFixedReferenceSweptAreaSolid geometry directly.
I fully understand that Safe wants to make the sending of Success or Failure emails in FME Flow Automations more granular. But the big issue for a lot of us users is that we are at a severe disadvantage due to limitations our IT faces with getting the whole SMTP info set up correctly in FME Flow Automations. For this reason alone, I think it would be useful for there to be a basic emailer that can provide us users with an option for actually being able to use FME Flow Automations and send the one thing we need … An email with success or failure info. As for our company (And I’m sure this is the case at a lot of companies) we can’t use FME Flow Automations and that’s a real waste considering we have a lot of uses for them. Sadly, the Emailer has become the major dealbreaker for us. It’s out of our control as users and wish there was a viable easy option to just have an emailer that can send without having to be tied to a company’s email server.
In an Automation email action, we would like to use deployment parameters to set the following parameters:SMTP Server SMTP Server Port SMTP Account SMTP Password Connection security (less important)Motivation: we have clients with an important number of automations that use the “Email (send)” external action. When the smtp server settings change, e.g. between DEV and PROD environments, it’s a significant amount of manual and error-prone work to change these settings consistently in all the automations. It would be a big help if these settings could be set using deployment parameters:
It would be nice to be able to use ArcGIS web connection to authenticate processes when using the ArcPy library
My organization recently migrated from Apache ActiveMQ to Apache Artemis due to ActiveMQ vulnerabilities and we had to migrate an automation to a different workflow since there is currently no support for Artemis.
Whenever one uses an SQLExecutor for inserting, updating and deleting rows (collectively known as DML statements), it is usually handy to know how many rows have actually been affected by that operation. I am by no means the first person to want to do that with FME.All DBMS I can think of have at least some way of exposing that information.Postgres returns a command status tag with the command type and affected row count, Oracle has the SQL%ROWCOUNT variable which countains the number of affected row in for the last query in a PL/SQL block, Snowflake stores and exposes query result stats on a per-statement basis, JDBC drivers have both the Statement.executeUpdate() and the Statement.getUpdateCount() methods returning the affected row count, ODBC drivers have the SQLRowCount functionMost notably, none of these are exposed in a way that an FME user could make use of, as they are part of some interface that extends beyond SQL statements. Most workarounds will involve a much slower operation, such as WITH (INSERT … RETURNING *) SELECT COUNT(*) or something similar.It would be nice if FME could return some output feature for DML SQL statements where the affected row count (potentially separated between updated, deleted and inserted, for databases that keep track of them separately, like Snowflake during an upsert).
Hi,the birds are chirping that Bentley is soon to phase out ProjectWise login via logical accounts. It would be nice if there was an option to connect to PW via IMS login as a transformer or an option to configure the connection in the settings
It would be great if we had the ability to set the orientation of the Junction objects in workbench. This would help when the direction of features were heading downward or even from right to left. At the moment I need to live with this kind of look (see image) and I wish i could just rotate the junction 180 degrees in this case.
Hi there, When using readers in a larger workspace, sometimes I might need to reference the reader (or a bookmark involving the reader after some post processing like changing coordinate system or renaming some attributes) multiple times in the workspace, either at the start, middle, or end. If something is being used, say 5 times, and crosses multiple bookmarks and transformers, I will use hidden connections to keep things readable, however it can still be a hassle if I need to use a certain reader/transformer/junction that is at a certain place in a large workspace, where I might currently be working in another area. I will then need to zoom out and scroll to it, and then drag the connection all the way over to where I want to use it. A way to get around this, is to have the ability to pin readers/transformers/junctions that are used a lot in a workspace and have many connections already, so that if it needs to be used again, you can simply connect to it wherever it is pinned in the window (and then can unpin it again after that if preferred), so that you can connect to any part of a workspace from any other location without having zoom in and out and scroll and repeat, just to create one connection Cheers!
Edit: Got the header of the post wrong but i cant change it. Should of course beTunnels: allow 1..N and N..1 connections Allow more efficient and cleaner workspace design with one-to-many connections for tunnels. Ideally if there is a 1..N or N..1 connection between transformers, one should be able to select all connections → right click → convert all connections to tunnels. The expected layout would be according to picture below. The N-side of the connection would only have one junction instead of N amount of junctions. Layout wise it is already achievable. It does however require many clicks as well as playing with alignment and placing tunnel junctions on top of each other. Therefore it is easy to snap connections by mistake.
I often want to view the source data is the native software and open it from FME Form via Open Containing Folder, then open that file. My suggestion is to add an option to open that file in the native Windows program, e.g. ‘View Source Data’ in native, or ‘Open in Native’ so you remove a step. If the file was a DWG it would auto-open in whatever Windows/user has assigned to open a DWG, e.g.AutoCAD 2027. If it’s a CSV, it opens in Excel (or Notepad++ for example) if that’s what the user has set. Thanks.
hi there, it could be a good idea to be able to add pinned annotations on a workbench for important messages or things to keep in mind for a workbench, so that no matter where you scroll in the workbench the pinned annotation will show in the exact same place. Ideally the annotations would be around the sides of the workspace frame
When using a Dynamic featureWriter with the PostGIS format you can handle the FeatureType geometry these ways :From first feature From schema definition Define one specific geometry typeNone of them are really satisfying for me :From first feature : if you have multiple datatypes you need to order your features or to coerce them leading to the second one From schema definition : no fme_geometry maps cleanly PostGIS geometry types (no support for multicurves or multisurfaces) Define one specific geometry type : not dynamicIt could be great to use a specific attribute : fme_native_geometry_type to define wich geometry type to use in a dynamic writer.This could be for any format (not just for PostGIS but I need it for PostGIS 😉) This could be populated by readers You could provide some transformers : to set this parameter depending on the feature geometry to set manually this parameter (choose format and then geometry type) to coerce geometries as defined by this parameter
Summary Annotations contain a lot of information, and therefore take up a lot of screen space. On the other hand, summary annotations dynamically display the settings of a transformer, which can be better than using static Annotation.Therefore it would be nice to have the possibility to select what information is shown in a Summary Annotation.This would have the benefit of showing only the settings the user deems relevant, while at the same time automatically updating this information when these settings are changed.
It would be nice to add tunnels to FME Flow Automations like you can with FME Form workspaces. It would help organize complex automations and make them fit better in the UI.
There’s various suggestions out there explaining how to get the number of features written during an automation:https://support.safe.com/hc/en-us/articles/25407494315789-FME-Flow-Get-the-Count-of-Features-Written-in-an-Automation However, these processes seem a bit complicated given that the count exists in the log for the workspace. Maybe I am wrong and it would be very complicated to access this number.For the “Run a Workspace” action, I would like to have “Total Features Written” included within Output Attribute → Success → Event Attribute:This would allow it to be utilized in downstream actions such as emailing.Also, you would be able to use it within custom attributes and that would eliminate the need for additional workspaces, automation writers, split-merge blocks and http requests.
When a Flow App is created there is an option to customize it by adding header, footer and icon images. Often this will be a company logo or something that is common to all Apps.Currently, the images for each App need to be uploaded individually.It would be great to be able to reference a file from a Flow Resources location to avoid repeated uploading of the same image(s) for each App.
The current customization options for FME Flow Apps is quite poor. FME Flow Apps have the potential to be a centralised focal point/application for organisational/enterprise level ETL; however, they currently allow for minimal flexibility and design when it comes to styling.The lack of flexibility and options for design improvements mean that even with greatest graphics/design eye in the world, FME Flow Apps are often left looking nothing more than a data dropzone, workspace runner or ‘Click to Download’ link.Expanding the options for image customization and editing, greater flexibility in text design, or incorporating HTML would greatly enhance the visual appeal of FME Flow Apps. This would increase visibility and improve perceptions to none FME users who are given access to a Flow App.Documentation could also be created in the form of a technical article on ‘Designing and Styling an FME Flow App for your Organization’ to support this process.
Currently in FME Flow Automations, when using the “Run Workspace” action, the job queue must be selected from a fixed list of predefined queues. There is no ability to dynamically select the job queue using a published parameter or automation attribute.This enhancement proposes adding support for parameter-driven job queue selection, allowing users to pass a value (e.g., from an automation attribute or published parameter, etc.) to determine which job queue is used at runtime.
In-depth courses for all FME skill levels.
Explore our extensive collection of how-to guides and tutorials.
Log an issue with us. Our support team is ready to help.
