Re: partition based upon timestamp column per day

[santhosh edukulla]

Thanks Rick for the response.

Note1:
The reason for partitioning are:
1. Query performance : We have reports which needs data between two dates. Other queries are there as well.
2. Maintenance: User can delete the data greater than a given data after certain date time
3. Insertion: Its a multi core multi processor machine , we learnt that partitioning can help insertion as well.

Note2:
We use load data infile to insert the records.
Again here we have a question: How many records per csv we can have for load data statement? We currently are maintaining 40000 records. Will adding more can increase the insertion rate?

Note3:
I was told that timestamp column based partition is more resource incentive compared to date\datetime? If yes, how to partition by day using timestamp column.

Thanks!
Santhosh