redundant unique index on entity-first approach

[chungwong]

Description

Discord Ref
With the new entity-first approach, we can add unique as below

#[sea_orm(unique, column_type = "Text")]
pub url: String,

I found it will generate two indexes, at least on Postgres

                                        Table "public.url"
   Column    |           Type           | Collation | Nullable |             Default
-------------+--------------------------+-----------+----------+----------------------------------
 id          | bigint                   |           | not null | generated by default as identity
 url         | text                     |           | not null |

Indexes:
    "url_pkey" PRIMARY KEY, btree (id)
    "idx-url-url" UNIQUE, btree (url)
    "url_url_key" UNIQUE CONSTRAINT, btree (url)

As you can see, it generated two indexes for UNIQUE property. I seeked help from the PG discord channel and asked if it is necessary and the adivce I got back was hinting it could be a mistake to have both indexes and UNIQUE CONSTRAINT is more "CORRECT", althought the difference is minimal.

Expected Behavior

It should use just UNIQUE CONSTRAINT

Actual Behavior

 "idx-url-url" UNIQUE, btree (url)

is redundant

Versions

sea-orm v2

[Huliiiiii]

Reason: We created an index once in create_table_from_entity and once in create_index_from_entity.