FME Form

I want to try openCV, but I get completely stuck with no real error message at all. I tried a lot of different approaches, but even the simplest doesn't work whatsoever. Workflow for the simplest:Use the openCV comandline to mark positive samples and generate an annotation file. To make it as simple as possible, I tried 1 positive sample, resulting in this annotations file: F:\\temp\\opencv\\positives\\0.png 1 45 35 38 63 Next I made a folder containing negatives with 20 negatives (tried .png and.tif) Next up is the samplepreparer: negatives: use existing (tred overwrite too) positives:number of samples: 1width/hight 200 and 200 (tried numorous different values) I allocate a negatives and vec file location and that's it. No spaces in anny of the folder structures. I get rejected features.  Storing feature(s) to FME feature store file `F:\\fmetemp\\wb-cache--PkosFd\\Main_RasterObjectDetectorSamplePrepar -1 19 fo 1 _lt_REJECTED_gt_ 0 9db0cf751f86d66944dfbb4c75f54d358bd7f17d.ffsup

Hi Everyone, I have succesfully managed to use a Custom Parameter to extract traits from an Ifc File with the GeometryPartExtractor. But when using the same said Custom Parameter as the Left Value in a Tester Transformer, no filtering takes place. Can one use CustomParameters instead of manually selecting the Attribute Values?  Test: After extracting the Trait, I can manually select the now Exposed Trait as an Input to use for my CustomParameter, but that defeats the point as the Custom Parameter is not really a parametric input.   Example File: I have shown three ways I attempted to use the custom parameter:Top Tester: Trying to use Custom Parameter for both exposing the Traits in the GeometryPropertyExtractor and in the first Tester: No success Middle Tester: Functions as wished, but there is a step that needs to be taken to select the correct Attribute Bottom Tester: Old school, just selecting the Attribute and correspong Parameter needed for the Tester.  Maybe I am using the Custom

Hello, My current workbench is outputting an incorrect amount of Polygons using the FeatureMerger Transformer and I think the issue is partly to do with the number of vertices that make up some of the polygons and the fact some double back?Or is it perhaps at the PointOnAreaOverlayer Transformers as they seem to be creating too many records as well? Hopefully my images capture in essence what I want to achieve and the issues I am encountering. Perhaps FeatureMerger is not the correct Transformer I need? Thank you in advance

Hi I am looking for someone who knows/has expierience how to transform GIS Tree cadastre/register Data into BIM Objects with simple geometric representation of Tree crown, habit, root ball and trunk and of course the register Information as Attribute Data attached?  If anyone has ever done such a project, or knows someone who knows someone. I would be grateful to get in touch to get some basic information about their Project. Greetings tobias

Hi, fellow FME-lovers! I have an interesting problem that you can help me solve in an efficient way. I have many pipe segements as solids in an ifc-file, see example in Test2.ifc. The pipes are in 3D and can be connected in all directions. (I'm not sure about the fittings but for now don't mind them.). Now, I need to calculate length per segment and I started working no a complicated scripts (2D and algebra...) but there must be more of an FME-way. I'm thinking along the lines of creating a center line from pipe start to end... If you have any ideas I will gladly listen.  Bonus: what if segments are bent? Will this be handled in you solution? Peter

So I have written out the points into each featureclass that overlay a polygon... but now I want to get a list of unique values of one particular attribute (with a total count value of each), for each polygon. So that means the unique list (and count) will be different for each polygon. I've tried this workflow using StatisticsCalculator to get total count of point attribute (that I want unique values for) and grouped by the same attribute. Then passed to AttributeCreator to create a new attribute with this value. Then passed to Aggregator and then fanned out to a writer. However, the values are the same. I need to calculate the total count of the unique values for each polygon. Please help.

I have path_windows for all dwgs in a folder. I could get AutoCAD file OR Civil3D files, I won't know that. How would I get results which I canIs it Civil3D or AutocadGet version for Civil3D. With the top autocad feature reader I figured out how to get the autocad version already, even when I have 2 files with a civil 3d features(these are now included in my tests but should be excluded from my result). I want to only get one result per file, autocad or civil AND the version per type. I tried many things but I cannot find a way to get "there are ciivl these are autocad" not all Civil has pipes network for example so the sum of civil types isnt a guarnateed solution.Current results: Top Feature Reader is AutoCADthe Bottom is Civil3D results I want   my current AutoCAD version results(the 2 marked read have pipe network so they should be marked as missing)

Hi, Is there a transformer that will take a table of OS Grid references of varying length/granularity and create a feature class of grid square polygons for them?

Hi there, so I have two column of data.​The column "Test_FloorNumber_Begin0" should have sequence of 0000, 0001, 0002, 0003 - the largest number of the input, or 0999. ​The second column "Test_FloorNumber_Begin1" should have sequence of 1000, 1001 - the largest number of the input, or 1999.​The sequence should start from 0000/1000 to the largest number from the input data. So I want to use FME to find out the missing number and output it. In this case 0000 to 0003 and 1002-1003 are missing.​Please help me as I want clueless on how to work this out.​Thank you very much.

I am using Bufferer to create a 4 foot buffer around right-of-way polygons. The tool works for 99 percent of the areas but I am getting a few random missing polygons as shown in the screen-shots below. I did find if I change the Buffer Distance one area will generate correctly; however, the issue occurs somewhere else. Areas: Buffer with missing polygons:Any Ideas how to fix this. Buffer works fine in ArcGIS for the same features.

Hi Community, I am trying Attribute reprojection transformer in FME to convert New Zealand Transverse Mercator Projection (NZTM) coordinates into WorldGeodetgic System 1984 (G1762) as per the attachment. However it is not showing my location of interest in my GIS software. Any suggestion on this. Regards, 

I am using the ACAD writer to convert GeoJSON to a CAD file. The issue I'm having is the ACAD Writer seems to be creating a bunch of empty layers named d_y_00, d_y_01, d_y_02 etc. that aren't in the GeoJSON I'm passing. The layers that are in the GeoJSON are written to the CAD file and work perfectly. The issues is I can't seem to figure out where these layers are coming from or how to remove them.

I have two FeatureMerger transformers fed by a single AttributeManager that outputs two string Attributes - one containing a GUID and the other a mixed-case name. The two FeatureMergers are identical, matching the GUID so that the corresponding name gets inserted into their respective Features. Oddly, one FeatureMerger converts the names to uppercase. So I did the obvious thing: blew it away and copied the properly-working FeatureMerger - only to see the copy convert to uppercase as well. The names enter (the Supplier port) in mixed-case and come out (the Merged port) as uppercase. The Requestor ports of both FeatureMergers are connected to the Merged ports of previous FeatureMergers, which themselves successfully merged two mixed-case Attributes into Features with (only) uppercase Attributes. So what is making FeatureMerger convert to uppercase? Edit: Replacing the "uppercase" FeatureMerger with a FeatureJoiner makes no difference. Nor does removing the AttributeManger's

Hi, does anyone know how to write input data to one "block?" I have input data - 8 frequency values (pic1), which are written to the dataset as individual "blocks" (pic2), but I need to write frequencies to one "block" ( pic3 ) Thanks a lot

I dynamically write tables (Soruce: PostGIS) with, as well as without geometry attributes into ESRI Shapes . The Writer for Esri Shapefile format generates entire shapefiles (shp, shx, prj, cpg, and dbf) even for tables without geometry attributes. However, these may not be  read correctly by ArcCatalog. For geometry-less tables, it would be better to have a dbf-file only! Is there any option for the Writer to avoid writing such entire shapefiles for geometry-less tables?

I have some polygons as streets (streetname) and many pipes which lay mostly inside the street polygons. Using the line on area overlayer I can get pipe attributes on the street polygons. But I would like to cut the polygons where one pipe ends and the next one starts.Another approach would be to cut polygons where pipe attributes change...is it possible?

I have a workspace which uses input user credentials to do stuff. I also want the workspace to work during testing by using System environment variables. This is working great for things like DBNAME, HOST, PORT etc where the user inputs are strings (text user parameters). I can simply just enter in ${PORT} for example and FME will look for a set environment variable. Where I'm having trouble is with the password. I can't do the same thing for the password so I've created a simple scripted parameter which first checks if there is a user password set and if not try and see if there is an environment variable. In most cases it works well, however, I've discovered that some special cahrachters are getting encoded by FME - So something like @ becomes '<at>'. Does anyone know a way to 'decode' these encoded strings. Ideally it'd be done in the scripted parameter, however, if it has to be done in the workspace so be it.Short from using some kind of regex in a string replacer I'm a bit s

Hello,is it possible to run from fme system call (or python caller) and don't wait for the result?

Hi I have a lot of small geotif tiles. I need to mosaic them together as geotif in specific zones.  I tried the standard compression which is jpeg. The problem is that some zones are to large for the compression method. I get this error: "GEOTIFF writer: Failed to write the output file using JPEG compression. The raster is being written with a tile size of '29005' rows by '82874' columns, whereas the maximum number of rows or columns that JPEG compression can handle is 65500. Please reduce the tile size if writing a tiled file, or reduce the size of the raster" So I'm thinking I only need to change the compression method. I tried all the other compression methods and I always get the same warning: "GEOTIFF writer: Existing value for creation option 'INTERLEAVE' was 'BAND'. Overwriting with the new value 'PIXEL'GEOTIFF writer: Existing value for creation option 'COMPRESS' was 'NONE'. Overwriting with the new valu

Hello,I am trying to populate a value into "_overlaps" using the PointOnAreaOverlayer Transformer but all the data keeps being Rejected.This works fine when both Readers are consuming Oracle Spatial but I think the issue may be because my Point data is non-spatial?Tried the 'Reprojector' Transformer but had no effect. The View with the Point data contains a SHAPE field therefore should contain geometry but the FME non-spatial reader does not contain the field and I cannot pick up the data using a spatial reader.I may be pursuing the wrong process here?!Thanks in advance. 

I am encountering an error where my clipped raster data is not being written to the designated output folder . I am using FME Workbench 2021.1. I was following the Raster Clipping workflow here. However, I must be making a mistake when adding the writer and working with the settings. Perhaps I was a bit confused as to how to set the writer and parameters. There error I received was something like FME file not accessible. Unfortunately, my FME crashed and closed before I could copy the error I received. I'm hoping someone can guide me through adding a writer. I am trying to clip ~3000 geotifs to a shapefile file of Ottawa, On. and fan out by Ward name. Thank you   

Hello All,I have an AutoCAD drawing that has small circles (polygons) sitting over the intersection of two or more lines. The lines finish at the edge of the circle but I need to extend each of the lines to the centre of the circle so I can then create polygon features from the lines. The resulting line from circle edge to centre needs to be merged with the original line i.e. not just create a new line segment.Any ideas would be very much appreciated.   

The thing I want to do is for example to when AGOSLAG(pic 1) = Kodvarde update AGOSLAG to the text Kodlang(pic 2). Then I want to do the same thing for MALKLASS, HKL. I have serveral tables in excel and they all start with the number 1.  I've tried with databasejoiner but could only update 1 attribute at the time there. I have about 20-30 attributes that need to be updated.  What is the best way to solve this problem?  best regardsJesper

OK I give up - help me!This is my first attempt to use the PythonCaller (since FTPCaller can't yet download directories or accept wildcards).  I've cobbled together some Python to retrieve a directory listing from an FTP site which I want to output (probably as a list) so I can generate the name of the file I want to download and pass it to FTPCaller.I've done a whole lot of Googling but I can't fix this error on the feature.setAttribute line:2016-10-12 17:46:48|   1.9|  0.1|WARN  |Python Exception <TypeError>: not all arguments converted during string formatting2016-10-12 17:46:48|   1.9|  0.0|WARN  |Traceback (most recent call last):  File "<string>", line 30, in inputTypeError: not all arguments converted during string formatting2016-10-12 17:46:48|   1.9|  0.0|ERROR |Error encountered while calling method `input'2016-10-12 17:46:48|   1.9|  0.0|FATAL |f_32(PythonFactory): PythonFactory failed to process featureimport fmeimport fmeobjectsimport ftplib# Template Function 

Basically I need to create new edit version from sde.default and need to compare one arcsde geodatabase layer(copy layer) to another layer (source layer). Any changes in the source layer will be written to the copy layer. Copy layer is a versioned layer and the source layer is not versioned.