Hi ​@deanhowell ,

If the point of the question was "How to replace space in an attribute value with underscore", I think StringReplacer would be a general solution.

Thanks ​@takashi that part I have figured out but just can't get that attribute as part of the layer name for the feature writer transformer 

Why not create a extra python scripted parameter $(_protallayernamecorrected) where spaces are replaced by underscores?

return FME_MacroValues['_protallayername'].replace(' ', '_')

Thanks ​@nielsgerrits I have never tried this but looks like a possible option. Cheers Dean

Just be sure the scripted parameter is under the original parameter because parameters are processed in the order from top to bottom, else _protallayernamecorrected can’t find _protallayername and it will fail.

​@deanhowell , you can use ParameterFetcher to assign the value of a user paremeter to a new attribute, and set the attribute to the feature type name parameter in the FeatureWriter.

Alternatively, you can also use the @ReplaceString function to set the required layer name to the feature type name parameter in the FeatureWriter. 

@ReplaceString($(_protalLayerName)," ",_,caseSensitive=TRUE)

Thanks again ​@takashi. I am using a string replacer to update the fme_feature_type and that is getting me a step closer, but because I need to use the dynamic schema (although maybe that is my issue for the update) to ArcGIS Portal, the only attribute I can use is fme_feature_type

It is now finding the right layer in portal but giving me a python error when trying to update

Does the fme_feature_type_name on the schema have the value with the spaces or with the underscores?

Thanks all for the suggestions and assistance. I ended up using a python caller earlier in the workbench to replace any spaces in the actual file names with underscores.

From

rest areas/rest areas.shp
rest areas/rest areas.dbf

To 

rest_areas/rest_areas.shp
rest_areas/rest_areas.dbf

The end result

• C:/Projects/AIDP/DataToProcess/rest_areas/rest_areas.shp

• C:/Projects/AIDP/DataToProcess/rest_areas/rest_areas.dbf

For those interested here is my python code:

import fmeobjects
import os

class renameFeature(object):
    def __init__(self):
        self.folder_map = {}

    def input(self, feature):
        original_path = os.path.normpath(feature.getAttribute('path_unix'))
        directory, filename = os.path.split(original_path)

        # Determine renamed directory (and store mapping)
        if directory not in self.folder_map:
            new_directory = directory.replace(" ", "_")

            if new_directory != directory:
                try:
                    if not os.path.exists(new_directory):
                        os.rename(directory, new_directory)
                    self.folder_map[directory] = new_directory
                except Exception as e:
                    feature.setAttribute('rename_dir_error', str(e))
                    self.folder_map[directory] = directory  # fallback
            else:
                self.folder_map[directory] = directory
        else:
            new_directory = self.folder_map[directory]

        # Now reconstruct real file path using renamed folder
        original_full_path = os.path.join(new_directory, filename)
        new_filename = filename.replace(" ", "_")
        new_path = os.path.join(new_directory, new_filename)

        try:
            if new_filename != filename:
                os.rename(original_full_path, new_path)
                feature.setAttribute('rename_status', 'renamed')
            else:
                feature.setAttribute('rename_status', 'unchanged')

            feature.setAttribute('_clean_filename', new_path)
        except Exception as e:
            feature.setAttribute('rename_error', str(e))
            feature.setAttribute('_clean_filename', original_full_path)  # fallback

        self.pyoutput(feature)

    def close(self):
        pass

Hi ​@deanhowell ,

Good to hear you got a solution, but I think there might be a simpler way without using Python script.

According to the original screenshot you posted, you have configured dynamic schema derived from schema feature, and set "fme_feature_type" to the Schema Definition Name parameter. I assume that you are using FeatureReader to read the source dataset and the schema feature output from the <Schema> port of FeatureReader as the schema source.

In that case, the value of attribute called "fme_feature_type_name" in the schema feature should match the schema definition name - i.e. the value of "fme_feature_type" in the data features. That is, I think, if you have modified the value of "fme_feature_type" in the data features - i.e. replaced space with underscore, you can just modify "fme_feature_type_name" in the schema feature as well as the same way, so that the dynamic schema definition could work as expected.

Thanks ​@takashi I will give that a go as well and see.