Re: does unique index do more than just being a constraint?

[Rick James]

Don't know about the docs, but a UNIQUE index is an index. Querying by the field(s) in the unique index will be (much) faster than if you did not have the UNIQUE index.

There is a slight difference between a UNIQUE index and a non-unique INDEX:
SELECT * FROM tbl WHERE xx = 1234;
needs to do only one probe if xx is known to be UNIQUE. If, instead, xx were a non-unique field, the query would start at the first occurrence of 1234 and scan forward until a different value is encountered.

Having both
UNIQUE(foo)
INDEX(foo)
is inefficient because both indexes have to be updated on INSERT.

In your case, DROP the non-unique INDEX.

All (with a few exceptions) indexes (including UNIQUE and PRIMARY KEY) are implemented as BTrees. This is a very efficient way for locating an individual row, or a "range" or rows with consecutive index values.

(The above comments apply to both MyISAM and InnoDB.)