Subject Changing COLLATION from UTF8 to UNICODE_CI_AI slows down select
Author Werner F. Bruhin
I am doing the following select:

SELECT NAME FROM CEPAGE_SYN WHERE UPPER(NAME) STARTING 'SEMILLON' ORDER
BY NAME;

PLAN (CEPAGE_SYN ORDER CEPAGE_SYN_IDX1)

Why is the above not using the index but below is?

SELECT NAME FROM CEPAGE_SYN WHERE NAME STARTING 'Semillon' ORDER BY NAME;

PLAN (CEPAGE_SYN ORDER CEPAGE_SYN_IDX1 INDEX (CEPAGE_SYN_IDX1))

With this it is using the index again:
SELECT NAME FROM CEPAGE_SYN WHERE NAME STARTING 'Semillon' collate
UNICODE_CI_AI ORDER BY NAME;

But upper/lower case of search criteria and column content have to match.

I would like to be able to search ignoring case and accent in both the
column and the search criteria.

Any hints would be very much appreciated.

Werner