Skip to main content

I am trying to get all the unique values from each field in a dataset. I would like to be able to expose the field name into each record, so I have an output table that looks like this:

 

image.pngI can run a statistics calculator to get the unique values, but the only way I have managed to expose the field name is by using an attribute exploder. This seems a bit messy because it outputs a whole bunch of combinations I am not interested in.

 

In the attribute manager, using @Value(fieldName) will return the actual value rather than the field name itself - are there any tricks using string expressions etc. to expose the actual field name? Seems like a simple thing to do but may be better handled in python. Just wanting to know if it's possible in FME! Would be really handy.

I think the AttributeExploder is your best bet still. I don't know how you can check for unique attribute values without breaking up each attribute into its own feature.

I tried with a SchemaScanner, that'll get your list of Field Names, and you can get your values from each field with a @Value(@Value(name))

imageBut fundamentally that's doing the same thing as an AttributeExploder at that point, it still results in every attribute with every value which you have to filter.


Hi @kaz​ ,

I agree that AttributeExploder is a quick way.

You can exclude attributes which you aren't interested in from exploding by setting a pattern (regular expression) of their names (e.g. "fme_" etc.) to the Ignore Attributes Containing parameter.

See the help of the transformer to learn more.


Reply