The Email column exported employees.email, which is the person's
private address. That does not belong in a file sent to an outside
survey provider, least of all as the address invitations go to in the
employer's name.
The column now stays as a placeholder for the work email, which does
not exist in the schema yet and will be added later. It is empty until
then, but keeps its place so the column mapping set up in Honestly does
not have to change once the address arrives. The export no longer
reads employees.email at all, so it cannot end up in another column
by accident either.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
A participant list for the employee survey on honestly.de, offered as
CSV and Excel right under the full data export. Columns: Personalnummer,
Email, Firstname, Last Name, Language, Location, OU, OU+1, ... , Role.
Language is always "de", Role always "Respondee".
The org columns run bottom-up: OU is the person's own unit, OU+1 the one
above, up to the top node. Their number follows the deepest chain in
the file, shorter chains are padded with empty cells, and there is
always at least one OU column so the file keeps its shape.
It exports exactly the people the full export would, with the same
filters. To guarantee that, the selection moved out of the full export's
route into lib/export-auswahl.ts and both routes use it; the full
export's columns are untouched.
Email is employees.email, which is the private address and optional --
there is no work address in the schema. Missing addresses stay empty
rather than getting a placeholder that would receive an invitation.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
