Index, add all, proper name

Question

I am creating index entries including a large number of proper names, and I want to Add All so that every occurrence of the name is indexed, but by last name.
I see the shortcut to create an index entry for a proper name, but in order to "Add All" I need to use the Index Panel.
To Add All for a proper name and have it appear in the index sorted by last name, I would use the Index Panel and then type the last name in the "Sort By" box (Ex.: for "John Doe" I would type "Doe" in the "Sort By" box). Right?

Answer 1

I don't think a single GREP could do it, which leads to another problem: how to prevent successive grep replaces from working again on an already reordered name?
Try something like this. Select the entire index; set the text colour to "Red". Then do as much as possible with GREP, and do the remaining by hand. The GREP Search should specify "Red" text, the replace should set the colour back to "Black".
Likeable GREP expressions are:
Search: (\u\l+) (\u\l+); replace with $2, $1. That'll take care of all "John Doe" to "Doe, John"
Search: (\u\l+ \u\.) (\u\l+); replace with $2, $1. That'll pick up "John A. Doe" to "Doe, John A."
Search: (\u\l+ \u\l+) (\u\l+); replace with $2, $1, changes all "John Allen Doe" to "Doe, John Allen" -- but for compound surnames ("Matt Le Blanc") you'll have to double-check...
Search: (\u\l+) (\u\l+\u\l+), replace with $2, $1, to alter "John McDoe" to "McDoe, John".
(I can combine all of the above into a single GREP, but for clarity I've restrained myself :-) -- besides, you probably need more than one S/R run anyway. Hence the colour trick.)
For slight variations, all you have use are these basic GREP wildcards: \u for uppercase character, \l for lowercase, '+' for one or more (hence "\u\l+" -- an uppercase followed by any positive number of lowercase). The period is a GREP wildcard, so it has to be escaped as "\.", and the parentheses define Replace groups (the $1, $2 in the replace; I put them around as "$2, $1" to switch first/last name).
The remaining should hopefully be exceptions to these simple rules, and probably could easy be done by hand, as they stand out in red.
>I will need to create one index entry for "Doe, John" which also identifies page numbers for "John Doe," "John A. Doe," etc., as well as a different index entry for "Mrs. John A. Doe," "Mrs. John Doe" and "Mary Doe" and the like
That's a bigger problem, actually. The page numbers for the variations of "John [X.] Doe" will all appear as individual entries in the index. They would have to be added
i all in the same way
as an index entry for ID to see they're the same. Perhaps you could first generate the index, look up strays, such as one entry for "J. Doe" and one for "John Doe", and correct these in the text. Only when the index entries are okay, do the final index cleanup. And hope the text doesn't get revised.