Set Data Type option within the AttributeManager

It would be so helpful to have this within the AttributeManager, instead of having to do this on every writer feature type (consider that there are many in there...). Besides, the transformer should preserve the input attribute field type and length alongside having the capability to define them for created fields.

Then, FME WB would become my favourite tool for remodelling ESRI GDB schemas. Right now, I'm going back to editing Workspace XML.

I agree with the previous poster. FME is reading the native data type format for the attributes populated from the reader so it should have no problem available in the AttributeManager.

Being able to define data type in the transformer is a better and more stable place to define the schema than at the writer stage. I often have workflows where there are multiple writers needed for various outputs.

Just wanted to point out that this is possible within an AttributeManager. Just use the Math functions in the "Attribute Value" field.

For example use:

to set the data type to a double (64 bit real).

also being able to define domain/subtype on geodatabase attribute manager fields

This would be a major help as it is one of those things I seem to have to do almost in every workspace, often for multiple feature types.

We've put in the foundation to do this, thanks for keeping it front of mind. Agree that it will be very useful.

Transformers creates additional attributes to my table. If I want these additional attributes in my writer, I see no option of copying attributes from transformer. I see only copy attributes from reader and others which I do not understand like manually, etc. Instead, I need to type each new attribute into the writer, set the data type (which I am not sure about) and drag in the attribute from the transformer to the writer. Is there any easy and safe way so I do not assign the wrong data type?

Sure. We need this option

for example, in Pro, when creating a field, we can specify its type

i tried the suggestion proposed by bleutzinn but appears not to work with me.

what could be the issue here?

Or better then this, have FME automatically change the type to the correct type (ie. if there's only numbers, then the output should be an integer or double depending on if there's a decimal or not, otherwise make it a varchar)...

This is in the latest 2023 betas (which will go live on or after April 13 2023) and I can advise that the AttributeManager now has an "Import->Data Types From Feature Cache" which will do a best fit data type scan with the outcome you describe above. It is a thing of beauty if I say so myself.

The way to get what you want is to set your writer feature type to "automatic" mode for the schema. Then the attributes that were added in transformers are just there. And if you use the all-new FME 2023 beta AttributeManager, you can ensure their types are what you want them to be.

Sadly that just affected the internal FME "transport" type, which is how we ferry the data from source to destination. It didn't actually affect how the output format would save the data. This is now made explicit in the FME 2023 AttributeManager.

Deeply agree.

Its only chages github, wiki codes to xlm, 64bse code…

We've updated our AttributeManager how-to article to utilize this new functionality in FME 2023.0! Check out Control Your Data's Attributes with the AttributeManager Transformer.

Mark and Dale also did a deep dive into Data Types and FME: FME 2023: What's Your Type? The AttributeManager Transformer Has the Answer!

great