Hi we are using version 11.2.0.4 of oracle exadata. We have a big table having ~5TB in size and holds ~85billion rows. Its having ~210 partitions. each partition holds ~500million records. We want to create new local index on one of the column which is varchar2(50) data type and having ~3million distinct records with 36billion null values. So we want to know how much time and resource(temp/sort space requirement) will it take in production and what will be the fastest way(online/offline) to achieve this. We don't have any other database with exact data volume, so we are planning to create few(around ~5 partitions) partitions with same data volume as in production. So wanted to know expert suggestion, if i can just prorate the resource usage estimate to ~210 partition and that would be the correct guess? or any other way should be followed. Is there any other thing which needs to be taken care which i am missing, apart from above mentioned resources?