Shape the future of FME with your ideas
Open ideas have been reviewed by our Product Management and are open for commenting and voting.
***Note from Migration:*** Original Title was: SchemaMapper - Change icon color to Red icon for invalid entries, and highlight the part is invalid It happened that the SchemaMapper is no longer working for me.After I thoroughly investigated, I found out that one of the Filter Features part of the Schema Mapper parameters is no longer valid (in fact the attributes got renamed in the source Schema Mapping file). If it visually has changed the SchemaMapper transformer icon toRed to indicate error, and highlighted which entry I should spot and validate, I wouldn't spent that much time to figure out what happened to cause the failure of the workspace.At least, I'd suggest something like what shown in the snapshot.
It would be great if the MeshSimplifier had an option to preserve edges.This would be very handy for example when you have multiple terrain tiles that you need to optimize but you want to make sure that the edges between the tiles meet up without gaps.
***Note from Migration:*** Original Title was: Restore: Enable to link Point Cloud XYZ Reader Component Mapping to a User Parameter In FME 2019 and earlier, we were able to create user parameters to allow users to specify component mapping for a Point Cloud XYZ reader appropriately according to their data schema. For example:Definition of the COMPONENT_MAPPING parameter (Python scripted):import fmexcol = fme.macroValues['X_COL']ycol = fme.macroValues['Y_COL']zcol = fme.macroValues['Z_COL']return 'x,col%s,Real64,y,col%s,Real64,z,col%s,Real64' % (xcol,ycol,zcol)However, I found this disabled in FME 2020.0 and 2020.1 beta, so some existing workspaces cannot be updated with FME 2020.I cannot understand why you did that. Strongly hope this will be restores.
The RasterDemGenerator produces a raster with an interpretation of REAL64. It would be nice if we could specify the interpretation in the transformer rather than needing a rasterBandInterpretationCoercer immediately thereafter.
Currently, FME Engines must be located in the same network as the FME Server install. Remote Engines would enable you to deploy the FME Engine in a different network to the FME Server core. The Remote Engines would communicate securely with the FME Server Core over the internet. This would enable the following workflows:Hybrid Deployments - Spread your FME Engines across multiple clouds, on-premises and the cloud, or within a cloud in different networks (e.g. public, private clouds)Regional Offices - Deploy the FME Server in one office region, and then deploy Remote FME Engines into other office regions next to their data.IoT Workflows - Deploy the Remote FME Engines on the edge directly next to the sensor.In short, it will enable you to move FME closer to the data without having to manage multiple FME Servers.
The GoogleDriveConnector transformer fails when attempting to download a file if the file with the same name exists already in the directory.I have seen this post: https://knowledge.safe.com/questions/100344/overwrite-option-in-google-drive-connector-in-down.html and that's a solution, however, it adds additional transformers to the workspace.It would be nice if the 'overwrite' option was available within the transformer.
Being able to run Windows containers with FME Engines on them in a Kubernetes cluster would enable users to read/write to Windows only formats in a K8S cluster.
Create a transformer that can be used to simplify complex meshes, so it would be nice if we had a transformer that could be used to create smooth meshes from simpler ones.
Hi.I'm trying to utilize the Workspace Reader to handle my almost 200+ repository of workspaces.I stumbled over a few issues with the Workspace Reader, that I want to propose as improvements."PublishedParameters" include both published and private parameters + parameters to embedded custom transformers. But it doesn't include any way of distinguishing between them. Please add an extra attribute for this."WriterParameters" does not include e.g. the MSSQL_SPATIAL parameters for executing statements before and after translation. Are they available elsewhere ? If not, please add.This is what I've found sofar.Cheers
Hello,When generating with the numericrasterizer, it uses the altitude values to define the raster pixel values. But an essentiel point is not configurable and understandable: how this transformer chooses the pixel value when several points are located at the same XY but nut at the same Z. (easy examples : a tree, a power line, a building, any vertical things or thing above another one).We don't know if the value taken by the raster is linked to the average of the points in the area (pixel size) or with the min, the max, randomly ...?I have noticed twice that points above the ground are not always taken into account and therefore do not appear on the raster, which makes it wrong because not all the elements are included. This is a bit of a pity for a transformer that is supposed to make a representation according to the elevation...It would be extremely useful (see indispensable) to be able to choose which value of z is used. Because this transformer works perfectly, it gives really nice results but it does not allow to have something reliable and rigorous.I hope that it would be possible for you to add a parameter something like "z value dertermination" with -lowest, -average and '-highest choices :)Regards,Arnaud
Add the functionality to Read and Write MIDI FIles in FME.
WebVR allows developers to publish interactive VR content on the web, which can then be accessed from any VR device with a web browser (including mobile VR, like Gear VR, Google Cardboard, Oculus Go & Oculus Quest). Introduce a WebVR / WebXR Writer.
Is it possible to try to reduce the memory footprint used when FME Server / Cloud is running idle? Ie have no jobs?The starter instance of FME Cloud seems to use 2,7GB RAM when "no work is done" - is it possible to squeeze this down by 10%? 40%? Might be a good internal hackathon.The reason for doing so would be to make more memory available for the tasks themselves and avoid these frequent issues when you hit the roof on the lower-memory computers:2020-04-22 17:02:43| 16.4| 0.0|INFORM|ResourceManager: Optimizing Memory Usage. Please wait... 2020-04-22 17:02:43| 16.5| 0.0|WARN |ResourceManager: Freed 683495 bytes of vector data (memory decreased by 1256384 bytes) 2020-04-22 17:02:43| 16.5| 0.0|WARN |ResourceManager: Freed 42565632 bytes of raster data (memory decreased by 42566352 bytes) 2020-04-22 17:02:43| 16.5| 0.0|WARN |After freeing memory: tcmalloc GetStats 2020-04-22 17:02:43| 16.5| 0.0|WARN |------------------------------------------------ MALLOC: 834246680 ( 795.6 MiB) Bytes in use by application MALLOC: + 0 ( 0.0 MiB) Bytes in page heap freelist MALLOC: + 28814680 ( 27.5 MiB) Bytes in central cache freelist MALLOC: + 10370640 ( 9.9 MiB) Bytes in transfer cache freelist MALLOC: + 18742336 ( 17.9 MiB) Bytes in thread cache freelists MALLOC: + 4796608 ( 4.6 MiB) Bytes in malloc metadata MALLOC: ------------ MALLOC: = 896970944 ( 855.4 MiB) Actual memory used (physical + swap) MALLOC: + 971071488 ( 926.1 MiB) Bytes released to OS (aka unmapped) MALLOC: ------------ MALLOC: = 1868042432 ( 1781.5 MiB) Virtual address space used MALLOC: MALLOC: 17172 Spans in use MALLOC: 3 Thread heaps in use MALLOC: 8192 Tcmalloc page size ------------------------------------------------ Call ReleaseFreeMemory() to release freelist memory to the OS (via madvise()).
Please add 3D Oriented Bounding Cube to BoundingBoxReplacer
***Note from Migration:*** Original Title was: Change Feature Joiner behavior so records with Null in the left join on field are not ignored Currently, if you use Feature Joiner to join two data sets and you set it to be a Left join the Left join On field is null, the record will NOT emerge from the joined port. This is contrary to what one would expect from a left hand join where ALL records from the left table should emerge from the port (eg as in ESRI, MS Access, SQL etc). This is dangerous behavior and will no doubt cause a huge number of errors.To get something that looks like an actual left join, you have to filter out the null values in the left key and add them back in later, which is utterly crazy in my opinion.Please could you either change the behavior so that a left join behaves as would be expected, change the nomenclature so that you don't use left, right join etc (the very nature of naming conventions the transformer makes it seem more SQL like) or at the very least have a warning to indicate it is unable to process null keys.
It would be very useful for @Timestamp function to return @Timestamp("^Y^m^d^H^M^S") by default
In Tester 2019 there was a UI change, which changed setting composite tests totally. I prefer the old way of one composite test clause, in which it's much easier to put opening and closing brackets and faster to define the test.Idea: Keep the old composite test. The new UI doesn't need to be changed, just add the old composite test syntax under the test clauses.
Hi,I have a .las file which I'm trying to convert it to .obj. I also have .dwg of the point cloud file. Since I'm new to FME, I was wondering if anyone here already has done this process and I can use their experience. Thanks for your help
Catches a translation error and outputs a feature containing an attribute that stores the last error message. If this transformer exists in the workspace (or a Bookmark?), a fatal error can be detected without stopping the translation, and the user can handle the error as a part of the workspace.Works like the 'catch' clause in a C++ program or the 'except' clause in a Python script, the user could improve the robustness of the workspace.The Terminator transformer should work like the 'throw' in C++ or 'raise' in Python if the 'ErrorCathcer' exists in the workspace.[Addition] It might be ideal if a Bookmark could work like the 'try' clause.
There isn't quick way to clip surfaces by areas currently. Perhaps introduce a new transformer which performs clipping surfaces by areas, or upgrade the Clipper transformer to accept surfaces as Clippee.It would be ideal if solids could also be clipped by areas.
I'm liking the new stats calculator setup in 2020, however i think the fact that you cant rename the output attribute inside it like the previous version is a bit limiting, meaning you have to add a subsequent transformer to handle that.
We often run long workspaces on secondary machines, and that means that floating licences are tied up unnecessarily once the workspace is complete, until someone checks the session and closes FME workbench.It would be nice to have an option to "save and close workpace on successful completion". That way the licences become available as soon as they are not needed, but if there was a failure we would still have access to the feature counts and inspection bubbles to help troubleshoot the issue.
Add support for FeatureWriter as a source for the zip file content of the data download. At the moment only Writers are supported.
The message should tell the user which version of FME the workspace was created with so that they know which version to use. At present I have to open the workspace in a text editor to determine the version.
It would be nice to be able to handle information about the lifecycle of Workbench I make. Not only attributes like version numbers , create dates etc., but also status like 'in development', 'ready for test', 'accepted', 'in production', 'end-of-life'. And it would be marvelous if FME Server would support this info too. E.g. that I can filter all WorkBenches that are not accepted, or that I can easily take all 'accepted' WorkBenches into production.
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.
