How to best perform delta of large tables in SQL as part of ELT data warehouse load?

  Problem: In our project there is a staging table , say 'STG_TABLEA', with 2 partitions, partition 1 for storing today's data ('DAY 0') and partition 2 for yesterday's data ( 'DAY -1' ) data. We would like to find the delta by comparing current day's data with previous day's data, i.e compare 'Day 0' and 'Day -1' data, and then extract new or modified records from STG_TABLEA. Other tech details: Database : Oracle 11g Extraction : to use SQL (no PLSQL or other ETL/ELT tools to be used) Each day would have around 30 Million rows, thus one would need to compare 30 million