Greetings,
I'm suggesting we replace:
> First, the column or row of Data_source that is at the index specified
> in the first value of Sort_index (name it "A" here) is sorted in the
> order specified in the first value of Sort_order.
>
> In case Sort_index has more than one value, next all columns or rows
> with the same value in column or row "A" are sorted in the column or
> row specified as the second value in Sort_index, in the order
> specified in the second value of Sort_order, and so on. If Sort_order
> has fewer values than Sort_index, the last value of Sort_order is used
> repeatedly.
With:
Columns or rows of a Data_source are sorted as specified by their
Sort_order.
Where multiple identical values appear in the first sorted row or
column, values aligned to those values create a subsort by its
Sort_order. Otherwise, values in sorted columns are sorted by their
Sort_order.
If no Sort_order is specified, it defaults to 1. If a Sort_order is
specified, then the last Sort_order specified is applied to all rows or
columns that follow.
Note: The behavior of the last Sort_order may change the default Sort_order.
********************************
Questions:
1) Is alignment of values limited to those in the first row/column?
2) Is subsort the right name for sorting values adjacent to the same
values in row/column one?
3) Do we need to define aligned to?
Comments and corrections appreciated!
Patrick