Table Performance with more than 83 million records

[esttevan]

I have a performance problem on a table that has more than 83,000,000 records. I need to make multiple SELECTs in it being that most of the most recent inserted data.

Does anyone have any idea to start solving the problem?

[diegolenhardt]

You can use HINTS, create indicates according to your parameters, partition the table,

Finally, there are many outings, but we can only give the best solution when we know where the bottleneck is,

Take the queries and manages the execution plan of them, which will show you what the fun's "match" to return, if there is full table scan, etc., which index is being used,

Begin by posting Querery's Execution Plans

[vitorleandro]

The best way to perform tuning is with partitioning for date range. In this way, only the filter partitions will be accessed. This, combined with local indexes, will make everything stay super fast.

See if your Oracle is Enterprise and if your table has a historical date that is rarely or never updated. Ex: transaction date.

Any questions, will post there!

[CarlosStetner]

Friend if they are only queries that you are performing once, no nothing will always be used.

You can create a smaller table only with the data you need, so you have liberality and ease to work with the table.

If it is a definitive solution, it is what the staff has already spoke, it has nowhere to run away.

[dr_gori]

It puts all possible restrictions, especially in the fields that you have indices, for it to restrict the lines.

And tb everything the people said there.