My immediate thought would be to put these all in a list that way you have one feature with lists for the attributes. Then you can remove duplicates from the lists and test to see which lists then only have 1 element (and that the element is null). This is pretty clunky, especially if you have a lot of columns, so I am hoping there is a better way and will keep thinking about it.

Thanks ​@liamfez I was able to identify those fields with all <null> values by following the steps explained by ​@nielsgerrits here: 

...the total number of attribute fields with all <null> are 31.

How can I drop those fields from the original data? I’m thinking about using some dynamic gdb writer, see below:

That looks good.

Yes I think a dynamic writer would be the way assuming no prior knowledge of which columns are entirely null. You could potentially use your attribute selection to make a schema and then have your original data align to that.

Thanks ​@liamfez for the feedback...yes I’m struggling to transpose the rows into columns and having the schema before writing the data into the gdb…

...if you know any trick to transpose the following table into columns please let me know! :)