Bugfix/1375 empty headings fallback

[i-just]

Description

When data contains an empty heading/key, the value for that column will be preselected under “Pagination URL” on the “Element” page and in the field mapping on the “Map” page.

• Google Sheet and CSV: those data types are 2 dimensional, so solved by adding blank_heading_<column_index_number> as the heading

• XML: doesn’t allow empty node names, so this problem doesn’t exist there.

• Atom and RSS: both use the same processing as XML.
• Json: because this can be deeply nested, I opted for throwing an error if empty keys are found.

Applicable to v4 and v5.

Related issues

#1375