The first one is invalid so the second one wins.
You should always run tests to see which one is faster in your circumstances.

>I need to select records from a Parent Table filtering on fields
>in the Child Table. But I only need to select (show) fields from the Parent table.
>
>I am deciding which approach would be faster:
>
>

>*-- Approach 1. Selecting records from both tables with JOIN and
>*-- then grouping them
>select MAX(PARENT.FLD1), MAX(PARENT.FLD2) from PARENT
>  left join CHILD where (&cChildFilter) GROUP BY PARENT.PK_FLD
>

>
>vs
>
>

>*-- Approach 2. Selecting record from Parent table with subquery
>*-- from Child table.
>select P.* FROM PARENT P WHERE EXIST (SELECT * FROM CHILD C WHERE
>(&cChildFilter) AND C.PK_FLD = P.PK_FLD)
>

>
>I would appreciate any suggestions/input.