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.
New Certification for FME Server
When I publish a workspace to FME Server and set the job success and job failure topics for the service I don't see the resulting publication listed in the notifications page (publications) of the web UI. It would be good to list these as workspace publications when this happens. This would also be great when a workspace is published and the Notification Service is set and topics are selected. This should also place an entry in the publications list.
Add all the key/value pairs of a list as attributes. Essentially the equivalent of @SupplyAttributes(&key,&value) for each element in a list.
Add the ability to generate tiles with overlap, either by number of pixels or by distance. Current hack is to get the bounding box, tile it, buffer the tiles, get their bounding box (ensure square corners), clone the image(as many times as there are tiles), and clip the image by the tiles with a group-by such that each tile clips the original image. (This last can be avoided if the clipper can deal with overlaps)
I need to group objects with the same classification if they are close to each other.But each class has its own value for the allowed distance.That value is stored in an attribute.I can group the features in the NeghborhoodAggregator, but can not use the attribute for width and height.I would like the option to use an attribute added to this transformer.
MapInfo supports per row symbology in all supported databases (Oracle, SQL Server, PostGIS, SQL Light).It would be nice if all these database readers and writers support MapInfo Symbology not just Oracle Spatial Object reader.Storing the per row style is the same in all databases as in Oracle: the symbology string (which is the same as in MIF format) is stored in a column usually called MI_STYLE.
With old FME versions, it was possible to use both lowercase and uppercase attributes when writing out to Oracle databases.This is no longer possible since FME 2014 build 14131. The only workaround I found is to copy+paste feature types from old workspaces to new ones. It is not really user friendly and I assume that in this manner the old 2014 Oracle writer is used even with FME 2016.
A transformer that would split each input polygon by any intersecting lines. Sort of like the inverse of LineOnAreaOverlayer. I can accomplish this using existing transformers, but it takes about a dozen transformers to do so (several of them blocking).
The current tool can output a lot of “sliver” holes or spikes if the data is not clean to begin with. The majority of these holes approach or are at zero area. (The hole vertices are essentially collinear.)
I think the 'Accumulation Mode' parameter in the Aggregator should be set to 'Merge Incoming Attributes' by default, when the "Mode" is set to "Attributes Only". No one noticed ever?
***Note from Migration:*** Original Title was: Better and more consistent support for lists/collections (of one) when reading XML Hi,I've just encountered a very irritating problem, that comes from FME handling lists/collections in an unexpected way (at least to me).When I read an XML file, each item may have a collection of sub items, which may contain 0 or more items.When FME reads the parent item, and finds sub items, it stores these in a list.EXCEPT when there's just one sub item! Then the sub item attributes gets stored in non-list multi-level attributes.This is inconsistent imho. If there are _any_ sub items present, a list should be always be generated. [MY SUGGESTION]In the current scenario, I need three transformers to handle the case:Test whether the multi-level attributes exist (single sub item), or the list exists (multiple sub items).Rename the attributes if a single sub item is found.Explode the list if multiple sub items are found.If the list was always generated if there are any sub items, I only need step 3.The multi-level attributes could still be created, if a single sub item is encountered, to keep backward compatability.CheersLars I.
As discussed in this thread, the PDF spec allows for layers to be locked, this means that when the file is opened in a PDF reader the user cannot toggle the visibility of those layers. It would make sense to put general information (in the case of a map: logo, legend, title, grid, map collar etc) on locked layers and only have the user control the map data layers.Right now FME doesn't support this, we have to do this in Acrobat after the file is created, but obviously that's not an option for high volume or automated processes.
***Note from Migration:*** Original Title was: List of created variables (through VariableSetter) as attribute input in VariableRetriever transformer
Could the Redirect To Inspector menu item be added to the toolbar as a default? Or alternatively, would it be possible for Workbench to remember my button layout between updates?
Dont necessarily want to create a parameter but a note in the parameter popup that is just free form text.
Quite frequently I need to delete all but one or a few readers or objects on a workspace. For example, when re-using only one chain on a previous workspace, or inspecting 20+ shapefiles for relevant objects.Then it would be good to select stuff to keep, then hit Invert selection and Delete to get rid of unwanted stuff.
I often have multiple transformers to handle a small function that are part of a larger component that is already within a bookmark. It'd be handy to be able to create one annotation bubble with a description of them small function. For example, create a name attribute by using an AttributeSplitter to split a string then set the name attribute value to one of the list elements using an AttributeCreator. Create an annotation "Create name attribute" and attach it to both the AttributeSplitter and AttributeCreator transformers.
Administration page can reference the obsolete notification (watch directory, schedulers...) or subprocess with option (kill old process)
I'm quite new to the docker and docker cloud.It would be really useful if you could provide a standard stack file for the docker cloud enviroment. It should be quite similar to the compose file:http://downloads.safe.com.s3.amazonaws.com/fme/beta/docker-compose.ymlStack file definition of docker cloud:https://docs.docker.com/docker-cloud/apps/stack-yaml-reference/
In the current version of the FME(2015.1.2.1), it is not able to adjust the order of the transformer's port, so the situation as the following figure, will produce a line of cross.If allow the user to adjust the order of the port, Will be able to reduce the occurrence of this situation.
Would like the option on the PointOnRasterValueExtractor to not interpolate. Just extract the value directly under the input point. This seems like it would be an easy one to add. Thanks for the consideration.
***Note from Migration:*** Original Title was: Automatically import of the possible attribute values if chosen Attribute to Filter by has values for coded domains (for AttributeFilter)
Enable the CartoDB Writer to create PK and BTREE indexes on attributes
CartoDB supports a limited number of data types - string | number | datetime | boolean. Currently when attributes on the CartoDB Writer are set to "automatic" the integer data types come across as strings. These then have to be fixed manually.
The Directory and File Pathnames (PATH) reader in conjunction with the File Copy writer is very useful to construct or rearrange the directory hierarchy in the disk system.The features from the PATH reader have several attributes storing file path components (full path, directory path, file name with extension, file name without extension, and extension), but the parent folder name is missing.C:/abc/xyz/parent/file.extOf course it's easy to extract the parent folder name from the full path or directory path using a transformer such as the FilenamePartExtractor, but I feel something redundant.Sometimes the parent folder name will be used to create a new attribute (e.g. the destination file name), so I think it would be more convenient if the PATH reader could retrieve the parent folder name (e.g. "path_foldername").
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.
