Attributes in a list transposed across as new columns in excel

It's possible but not particularly straightforward in current versions of FME. Not sure whether the same sort of workflow would work in 2015

Potentially, but I may have to change things further down the line & we already have a script set up to work with such a ragged edge table. Perhaps we need to make some changes in that regards though.

Now I know I could do that with a ListBuilder/ListConcatenator, and it seems it'd be wise for me to use a ListElementCounter to get an idea of how many columns I would need.

Anyhow, I've built a very non-elegant solution. I'm just building a list and setting up an AttributeCreator to output a new attribute for each list element number.

e.g., AttributeName = Secondary_ID_1, Value = @Value(_list{0}.Secondary_ID)

AttributeName = Secondary_ID_2, Value = @Value(_list{1}.Secondary_ID)

AttributeName = Secondary_ID_3, Value = @Value(_list{2}.Secondary_ID)

...and so on until I've created enough new attributes to hold the max count from the ListElementCounter.

Perhaps a more elegant solution will come to mind & if so I'll post here.

Thanks all for the quick feedback.

import fme
import fmeobjects

def create_schema(feature):
    attrlist = feature.getAttribute('_list{}.Sec_ID_Name')
    feature.setAttribute('attribute{'+str(0)+'}.name',"ID")
    feature.setAttribute('attribute{'+str(0)+'}.fme_data_type','fme_char(20)')
    for i, val in enumerate(attrlist):
        feature.setAttribute('attribute{'+str(i+1)+'}.name',val)
        feature.setAttribute('attribute{'+str(i+1)+'}.fme_data_type','fme_char(20)')

Thanks again.

here is result

@takashi

Of course that's possible.

Just not very attractive because then you can't import a attribute names . Unless you create a mapping first and import that, which kind of defeats the purpose as it is clearly less efficient.

I simply create a mapping for this type of enumerated attribute name in say excell, making it very as large as anticipated need.

When not enumerated I just copy the list from the feature information window and parse it using like notepad++.

You could replace the exposer with python attributecreation ass the data is enumerated.

I guess doing that would increase the dataset by 2.

This will go on till safe figure out how to blindly expose everything hidden.

@egomm

So I can dynamically write dynamically created attributes?

How would the schema be created?

Always keen to learn some new technique, I am.

@egomm

To get to the attributes I would have to read the file again and still not know what the attributes are, unless I sacrifice dynamism.

Exposure of enumerated attribute names makes it possible to access them in the workbench. (I simply create a long import list)

Not required for the solution of this topic I admit.

But for this topic probably adequate.