Transformers

I have two tables: Room and Furniture. In the Room table, I have room_id, and furnitures attributes, and in the Furiture table, I have room_id and furniture_name...etc. How to link these two tables through room_id, so each room will have a list of furnitures that belong to that room? Thanks for your input!

I have a workspace that has as outputs some 3d models and uploads them onto Cesium. The inputs are a single band raster with elevation as a value. I convert this raster in a TIN and using appearance setter I drape a texture on top of it.Said texture is a 3 band (RGB) raster which is created from the elevation values, matched with a file that tells the ranges of height and the colours to apply. At the moment I generate this raster in GDAL using gdaldem color relief.  Does FME have a similar transformer or workflow?

I have been puzzling for a while how to process data available on the Dutch portal www.pdok.nlAfter downloading a gml-file (type Dutch TOP 10 GML), I would like to extract the information stored in elements and convert this information into attributes. Therefore I use the XMLQueryExtractor. This does the job, information is temporarily stored in lists but this is a very timeconsuming process. Examples of elements with multiple values from the featuretype ‘Wegdeel’ are naam, typeWeg ,hoofdverkeersgebruik and fysiekVoorkomen.  Is there a more efficient method? I have attatched an example workspace… Many thanks in advance. (v 2019.2.2)

I have two tables: Room and Furniture. In the Room table, I have room_id, and furnitures attributes, and in the Furiture table, I have room_id and furniture_name...etc. How to link these two tables through room_id, so each room will have a list of furnitures that belong to that room?

Hello, I wrote a tool to join 3 Excel spreadsheets (solar data and 2 different meter tables) and an address point feature class. I tried appending this data to a SDE geodatabase feature class called, Distributed Generation. The writer only appended the attributes and not the geometry. I'm not sure why the geometry is getting lost during the translation. (As a temporary work around... i wrote it to a file geodatabase and am manually appending in ArcMap).  I have attached a screenshot of my workflow and parameters of the writer. Any help or ideas would be greatly appreciated 😊  

In my data, I have 3 Oracle tables (numbered for ease of explanation); 1) a spatial table with geometries of tubes used as casings for HV-cables 2) a spatial table with the geometries of the actual HV-cables3) a non-spatial table with the connection between the tubes and the cables they contain (essentially displaying which key(s) in table 1 is(/are) matched to which key(s) in table 2 What I'm interested in is isolating the tubes which have a cable running through (/on top of) them based on a spatial analysis, but aren't yet connected in table 3. Therefore, I'm looking to exclude all tubes with a key present in table 3 from table 1 before progressing with the other translations (which are going well), but I'm not getting anywhere with Matcher. Is there any other transformer to use here?

Hello,I need to publish SOAP service to receive data from an external application based on a predefined wsdl. I found samples on using REST but nothing about SOAP. Where can I found some samples or hints to start working on this?ThanksRoberto

I have a very simple question, I have two line datasets that cross each other. The goal is to get a count of how many times each line in dataset 1 crosses a line in dataset 2 (total crossing, ie if line 1 crosses line 2 twice, would like to track that). To do this, I was using a spatial relator with the attached parameters. The problem is that the "related suppliers count attribute" is 0 across the board. I have used "Requester Crosses Supplied" and "Requester Intersects Supplier" with no success. Is there a parameter I am missing?

Hi,Im trying to create a best fit line from a list of coordinates on a runway and dont really know how to approach it in FME. Im familiar how to do a linear regression via the R caller but cant figure out how to create a line from the values. Im not even sure that a linear regression is the proper method to do collinearity. Anyone have an idea?Thanks

Hello!I need to get the properties of a big GeoTiff ( 146299 cols x 213475 rows or even bigger)Among these properties I need to extract the ref system, the resolution, splot the bands etc. Apparently, in order to get them FME needs to read all the raster and this process takes forever. Is there any alternative way to get these without reading all the raster?ThanksOskar         

Hi, everyone​I wanna export attributes to a word table, so I use MSWordStyler transformer to generate a table element before writing into a Word writer. The pic below shows how MSWordStyler's parameters are set.​ ​However, when I run the workspace I got a error says that "Error message: 'AttributeError: 'unicode' object has no attribute 'append'".​Is there something I miss or somthing set incorrectly? What should I do to deal with it?💗​​

How add a group by option in my custom transformer? Example:  Another example might be if I put a Counter with local scope starting at 0 inside the custom transformer, then I would like to group by for instance _creation_instance - to get three features out all with _count = 0. In the Custom transformer I see a setting Group by <not set> but I am not able to edit that one (there are no suggestions when opening it). I tried creating a parameter called fme_group_by but that did not work either. How can I do this?

Hi, I'm using a feature reader to read a bunch of CSV's with changing column and data row numbers. I first extract the row numbers with a text file reader en feed these values into a feature reader. Like so: When running the workspace it looks like FME interprets my expression as a text:Using FME 2020.1 here. Does anyone know what is wrong? The feature reader only allows for arithmetic expressions so no option to enter text in the "Data Start Line" field Freek

I'm newbie...I adapted the code of FME's exampleFME Server Playground - Upload File In Sessionto run as custom-widget within ESRI Web AppBuilder (Developer Edition) and I have unexpected problem. This problem occurrs also if I run the exact code of the original FME's example, since the code is good, only my cloud workspace apparently lack some definitions. I'll explain - If I run FME's original code example using FME's original easyTranslator.fme workspace, then for different input files I get different results (well, it's obvious expected behaviour...).Same if I run it on easyTranslator.fme which FME has put inside "Sample" folder in my cloud (I assume it's exactly same workspace). But when I run SAME exact FME's code example on MY FME's cloud workspace that was created independently, it returns ALWAYS the results zip code of the FIRST input file, no matter if I upload other file and run on it. And if I run it from the beginning on a n

Hello, I have two shapefiles :an existing polygon shapefilea newer file containing a line (more precise) that covers part of the existing polygon fileMy goal is to obtain a polygon in which the more recent line is replaced in the initial polygon. I have identified the indexes of the points that need to be deleted in the initial file but I don't know how to add between these two indexes the points of the more recent line.I thought of making a loop in a customTransformer but it seems quite complex... Do you have any suggestions ?

HiI am having trouble speeding up the ShortestPathFinder. The input is about 7 million from-to-lines and I was hoping to speed up the calculations by setting the "Parallel Processing" to medium on my 8 core system. According to https://docs.safe.com/fme/2019.0/html/FME_Desktop_Documentation/FME_Transformers/parallel_processing/parallel_processing.htm , that shoud render the exact number of cores. Still, I don´t notice any speed improvements and the Windows task manager reports about 6% CPU load. At http://docs.safe.com/fme/2018.1/html/FME_Desktop_Documentation/FME_Transformers_HelpPane/Transformers/shortestpathfinder.htm , it says "If it is enabled, a process will be launched for each group specified by the Group By parameter." However, the most recent documentation, http://docs.safe.com/fme/2019.0/html/FME_Desktop_Documentation/FME_Transformers/Transformers/shortestpathfinder.htm , does not mention parallel processing at all. Am I missing something? How do I enable

I am finding that ShortestPathFinder in FME Desktop 2019.1.0 gets exponentially slower as I increase the number of "From-To" lines, whereas I would expect it to get linearly slower (i.e. twice as many From-To lines, twice the time) as each path calculation should be independent. Can someone explain why this would be and how I can avoid it?To give some context, I am running ShortestPathFinder over a (road) network with simple 2-point "From-To" lines whose vertices have already been snapped to the (road) network. There are (up to) 120,000 from-to lines, with a network of 280,000 edges (as output by a LineOnLineOverlayer to split all lines at intersections). For ShortestPathFinder I am using default settings, other than turning off snapping (reducing iterations from 10000 to 1000 made little difference). [Strangely, setting "Re-order From-To Line" to "No" hides all the other optimization settings - a bug?]The final output results are all exactly as I would expect, but it takes about a day

I have tried using VectorOnRasterOverlayer, Clipper, RasterBandNodataRemover and RasterClipOverlay...all being fed to an ECW Writer. And even though the results are all different, no method is giving me a clipped (i.e. no black or white boundary) ECW version of our orthophtoto: To me my method appears to be logical, but I'm obviously missing something.  RasterInterpretationCoercer parameters: RasterClipOverlay parameters: Clip polygon parameters: ECW Writer parameters:  As mentioned, I'm pretty sure I'm doing something wrong (I'm also definitely not a power FME user), but what I'm trying to do shouldn't be that difficult either. The original orthophoto looks like this: I made the background grey so it is obvious there is no white or black data surrounding the image. The clip (or mask) polygon is just something I had used to clean up the original image, and it worked well for that: The best ECW I have managed to create so far is this:  But how do I get rid of the

I'm successfully using the ImageRasterizer transformer to create images of extracted parts of a 3D model. The result is a png with a "wired optic" (see attachment). Is it possible to fill the empty spaces with color in order to get a continuous, solid colored surface? Without having any empty spaces like I'm achieving now? Many thanks for your help 😃 Regards,vill

I am using Kafka connector in Batch mode to read and process 10000 messages in a given batch execution. I want to make sure Kafka connector commits offset for 10000 messages. In case of any failure in the job, and If I restart the job (Using EARLIER option), I want Kafka connector to read messages from next offset of last successful batch run. Please advise.

Hello there, Im using the Shortest path finder transformer to analyse best routes from A 's to B's where the network is a road system. There is however a bit of flexibility in my project where the route could sometimes go "offroad" to skip going long detours. So i used my MS Paint superskills to explain what I mean.  Lets say that the black lines in the road system (network) and the yellow line represent the best result from ShortestPathFinder for getting from A to B. But it needs to take that tedious detour to west and then east again and i would want to allow it to take the green dotted detour instead (marked X), even if there is no road present there.  Pretty sure that the ShortestPathFinder does not have that option so I was wondering if anyone here has a suggestion how to solve this? And then in a way that you can set a snapping tolerance requirement for shortcuts (for example has to be max 10m a part for it to be allowed) or it wont happen. And in a way you ca

I have a folder containing a number of esri lyr files which are used to render data in ArcMap, I would like to return the name of the Esri Featureclass from the lyr file, I have attempted to do this by using both the Esri ArcGIS Layer reader and the Schema Reader and attempted to expose the ARCGIS_LAYER_DATASET using the "AttributeExposer" but it only returns <missing> .Any assistance with this is greatly appreciated. Many thanks

Hi,Which combination of FME transformers would I need to use for splitting an irregular polygon into convex partitions (and possibly also minimizing the number of resulting parts). The following figure illustrates what I am trying to obtain:I have come up with an approach based on the Triangulator which can be described as follows:Sorting all interior lines from the triangulation by decreasing lengthConsider the first occurrenceIdentify the 2 adjacent polygonsDissolve those 2 polygonsCheck convexity of resultIf convex, keep resultIf concave, discard resultThe process should for instance escape the loop when all interior lines have been consideredI created an FME workspace which does this.The initial figure:would become like this after triangulation:And ultimately the end result would be the following where all parts are convex:However the iterations are hard-coded for the moment because the custom transformer I tried to create contains blocking transformers and multiple loops which are

Hi, I've created a workspace that reads multiple dwg-files, goes through some cookie cutting and outputs the results to a pdf. The cookie cutter is dependant on a variable. I've managed to setup a second workspace that runs the above for each feature in a dataset. It all works. The problem is that the dwg-files are beeing read every time which takes some time. So my question is, is there a way to read the dwg-files in the second workspace and just pass them through the workspacerunner? Tried for some time now and can't figure it out. Thankful for any advice!

Hi All,  I'm looking for some ideas on how to get the z value from a point cloud onto a point in porformant way. As far as I can tell there is no easy way to do it. Am i missing something? Options I can see available:  Convert to Raster with RasterDEMGenerator and use PointOnRasterValueExtractor - pretty cpu and memory intensive Convert to points and use NeighborFinder - pretty cpu and memory intensive Has anyone got this process figured out?