recognize column based on the existing content in the column

Hello

I am new here and have not found a solution to my problem, but it is probably obvious for experienced users.

My students have created several surveys and now I would like to get the received mail addresses from the received tables (xlsx). Unfortunately, these surveys were all individually designed and the columns are named differently or not at all. So I can only recognize by the format of the content (xyz@mail.com) whether there are mail addresses in a column (using regex) or not.

How can I use a node to recognize the desired column based on the existing content and then only use this column in the next node?