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.
There are several transformers that produce fixed attributes (NeighborFinder, TopologyBuilder, SpatialRelator, etc). When using them in a custom transformer I have no idea if I am overwriting any previously existing attributes and would like the option to specify a prefix for these attributes. Our local rule of thumb is all interim attributes in a custom transformer start with _!, and that pattern is never used outside of custom transformers. So for the NeighborFinder, with a prefix of _! would produce _!_distance, _!_angle etc.
I love the new XML elements to match capabilities of FME 2016 (still on beta), was always an hassle to get the right element. It would be even better if after scanning the XML document and displaying the scan results that a little export (to txt) option would be made available.
Using Quick Add by specifying the capitalised letters associates the combination you entered with the Transformer chosen. e.g. entering 'AC' and selecting AngularityCalculator, will select that Transformer in the list next time you enter 'AC'. But I don't use the AngularityCalculator often, so next time I use 'AC' to find the AttributeCreator I have to go scrolling for it in the list. It could be nice if they were grouped together so FME remembered all the 'AC' typed/selected transformers and grouped them together, then listed the remaining Transformers after. Maybe an FME Option to reset/edit the list or toggle this mode with the existing mode.Of course I could use my own shortcodes such as 'CA' for AngularityCalculator but that's harder to remember!
Within the history of FME workbench, it would be nice if you can add a mark to a step in the recorded history. This way it becomes easier to go back to a certain moment if you are testing different possibilities. Now you really have to search to find the step you want to reset, this way you can say: go back to that mark.
When manipulating a spatial database using SQL statements, there are many cases where the syntax requires SRID (Spatial Reference System Identifier). If you are going to use such SQL statements with the SQLExecutor, will have to find a proper SRID from a documentation. It might be convenient if there was a transformer which can be used to interpret any FME Coordinate System to the corresponding SRID.
In addition to being able to provide a filename filter (e.g. *.dwg) on a file-type published parameter, it would be great to be able to restrict the user to a specified directory via a directory filter.
Adjust database(-like-)writers in such a way that you will have three input ports: insert, update and delete. Within the properties dialog box, you should have radio buttons to turn the different writer modes on and off.Example: http://www.larsdevries.com/img/database%20like%20...
Description of action is required in undo tool to understand easily what action we are going to undo and we can decide whether it is required or not.
For debugging and testing (32bit vs 64bit and also version/build number vs version) and in a couple of instances for customers with >1 installation of FME Desktop, it would have been handy to allow the WorkspaceRunner to allow you to determine which FME the sub process should run in rather the one defined by the path environment variable. An option to select or detect which FME to run would be useful.In the past I have solved this by effectively building a Custom WorkspaceRunner by using a SystemCaller to do the job instead.
Enable user to define SRID to save into SDO_GEOMETRY when inserting data into existing Oracle Spatial table, even after truncating old data. This procedure is repeatedly used in transforming data between systems in and out of the company. The target table must stay the same as it is used elsewhere - if source changes it's alright for transformation to fail (of course we have sanity checks before running it :-).
Add the import options for Named database connections to the Choice (with alias) User parameters.This is already available for Formats and Coordinate Systems.
Any hope of getting a way to write 3D output directly to BabylonJS, a webgl driven javascript 3D game engine (physics, interaction, 3D rendering etc.) Could be a great way to push out 3D city-size models of areas (like what ArcGIS does with CityEngine (http://www.esri.com/software/cityengine) to explore 3D city environnents
Itwould be useful to rotate results in inspector as sometimes we are given mapsthat contain previous results and the only way to compare them is by twistingyour head so that you can compare them—the new FME Output with an old map—side byside. It might seem like a silly request but it can be very useful. Sometimes itjust makes sense to review the data by zooming in, rotating it to put thingsinto the right context.
***Note from Migration:*** Original Title was: Add a transformer like the AttributeSplitter that allows attributes to be split based on a delimeter or on a start and end index into multiple new attributes. Example:Input attribute TEST value 1,2,3,4Split based on delimiter: ,Output attributes TEST1, TEST2, TEST3, TEST4Example2:Input attribute TEST value 1,2,3,4Split based on index:start 0 end 2, new attribute TEST1 value 1,2start 3 end 6, new attribute TEST2 value ,3,4This can be done, but requires multiple transformers insequence.
In the FeatureMerger you can have the following combination:- Feature Merge Type = Attributes Only - Process Duplicate Suppliers = YesWith the scenario above, the Geometry Merge Type should not become active, since it is only allowed to merge attributes. However it does become active (and probably doesn't work) and it confuses me everytime.
I am doing more and more with Javascript frameworks (bablylonJs, a 3D webgl game framework, KendoUI Javascript controls, etc.) The nice thing about their documetation is that they not only describe the api, but also have a online playground (like DOJO, JSBIN, JSFiddle, etc.) where you can see how feature works.Here is my brainstorm. For some of the more confusing transformers, eg. FeatureMerger, what about a limited FME cloud driven space which gives you a JSBIn like framework where one can see an example of how the transformer works? So that it does not cost you much $$ to leave up. It could be restricted to a very limited data set, and only be one transformer linked to a "online version" of the Data Inspector. All the levers I would have would be to change the parameters of the transformer from the default demo to different values so I could play with it?
Give a count on every unique value inside an attribute would be extremely useful to improve cleanup of data flows. A simply popup after a right click on the attribute would be such an improvement and reduces the development cycle.
I need a GroupBy function in the ListRangeExtractor transformer. Now the complete range is extracted, but I need it to be extracted per group.
Managing multiple FME Cloud instances (and I guess this could apply to FME Server too) could be a bit easier if the instance name is mentioned in the page title, so that it also shows up as the tab title in the browser.
I created a Python Creator to export some features from Oracle Database using this python library cx_Oracle. And for each test, Python is more faster than Oracle reader with a where clause.43 features.Python results: 4.0 seconds to export Oracle to Json fileOracle results:6.1 secondsMay be Oracle Reader loses time in the initialization.
Netty is a great tool but Nodejs is very powerful. Do you have any plan to integrate a node server with FME Server ? :)
Sometimes FMEJobSubmitter lost the connection to FME Server because there is a time out or an another IT reason.Add an option where the transformer retry x times to connect to the workbench (example: Recovery mode).
Ability to select multiple attributes in StringPairReplacer (like StringReplacer). So don't create a result attribute or propose a table with source attributes name and destination attributes name.
I would like to throw any file at a (Feature)Reader, no matter what the extension is. So whether the extension is something weird or it's missing, if I have a KML Reader for instance, I would like to have the file opened as a KML (and if that fails, try it as a KMZ etc.).I realize this would require some additional checks for the Reader (check if it's zipped and so on), but I don't think this is too hard to implement?
The number of regular transformers is now more than 450. This allows users to perform various translations depending on requirement. On the other hand, this also brings some difficulty when searching a required transformer, especially to beginners.To reduce user's effort to search a transformer, how about adding a special folder (e.g. "Favorites") to the Transformer Gallery?- The folder contains most frequently used and/or most fundamental transformers (30-50) by default.- The user can add/remove any transformer to/from the folder optionally through easy operations on the Gallery interface. Maybe right-click menu commands are preferable.- The default transformers set can be restored. FME Options | Transformers: Click a button -> Show a message "Are you sure...?" [OK] -> Restore the default.- The folder can also be hidden optionally. FME Options | Transformers: Uncheck a checkbox e.g. "Show Favorites Folder in Transformer Gallery".The folder may be useless to users who are familiar with most transformers, but I think it would help beginners to make things easier.
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.
