[chaley]

You don't say if #readstatus can have values other than 'Read' or 'Unread'. I assume it can't. The templates below won't work if my assumption is wrong.

These two templates return 'Yes' if #readstatus has more than one value for a series, the empty string otherwise. The GPM template is probably easier for you to maintain. The Python template is much faster.

Edit: add comments to the template, and remove the trailing comma

program:
	globals(answers = '');
	if !answers then
# Get a list of series names
		series_names = book_values('series', 'id:true', ',', 1);
# Loop over the series names
		for series in series_names:
# Get all the values in the '#readstatus column
			values = book_values('#readstatus', 'series:="' & series & '"', ',', 1);
			if list_count(values, ',') > 1 then
# More than one value found. Store the series in the answers global for later use
				answers = answers & series & ','
			fi
		rof;
# remove the trailing comma
		answers = re(answers, ',$', '');
		set_globals(answers)
	fi;
	str_in_list(answers, ',', $series, 'Yes', '')

Edit: Fix bug where identical values were counted as unique

python:
def evaluate(book, context):
	db = context.db.new_api
	answer = context.globals.get('answer', None)
	if answer is None:
		answer = set()
		all_series = db.all_field_names('series')
		for series in all_series:
			series_id = db.get_item_id('series', series)
			books_in_series = db.books_for_field('series', series_id)
			vals = db.all_field_for('#readstatus', books_in_series)
			unique_vals = {v for v in vals.values()}
			if len(unique_vals) > 1:
				answer.add(series)
		context.globals['answer'] = answer
	return 'Yes' if book.series in answer else ''

The 'globals' stuff is there so the template computes the answer data once for the search instead of for every book.