Duplicated columns when index already has primary-keys columns

[ddorian43]

Hi,

So, assuming we have a table with `primary_key(a,b)`, and we create an `index(b,a)`.
Is the end index in key-value stored as `idx/b/a/a/b` or `idx/b/a` ?

Meaning, you already store the primary-key columns, so do you repeat them again or tidb is smart enough to not repeat them ?

Regards

[Ewan Chou]

Hi,

They are two separate indices, the primary key is `a/b`, the other one is `b/a`.