Sequential selection from DB

[Former Member]

Hi Experts,

I have a DB with more than 30 million records.

I would like to process the records using the packet size <p_pckt>.

My Requirement is :

I would first process the first <p_pckt> records first.

Then later on when I come to execute the report some other day i would start the fetching and processing from <p_pckt> + 1 record in the DB and so on.

Could any one suggest some efficient way of doing it.

Regards

Sandeep

[ThomasZloch]

a Selectus Interruptus, very interesting.

You could add a flag to the DB table (is it just one?) and create a secondary index on this flag. Then

- select records with emtpy flag and addition UP TO <p_pckt> ROWS

- process the data

- set the flag in the processed records to "X"

This assumes that there are no blocks of data that must be processed together.

Thomas