Which is the quickest way to blow up 64 million records?

[rrojo7229]

Hi,

Could you tell me please, which is the quickest way to blow up a table which has around 64 million records?

- Delete table or make a delete by record - (using for each) with -i (Crash Recovery option DISABLED) in the proserve?

Regards,

 

[taqvia]

does this table exists in seperate area if yes then best way is to truncate the area.

drop the entire table abd recreate it but it will require a compilation as CRC will change.

for each <table>:
delete table.
end. will be fast enough and also the bi size will also not grow as transaction scope will be for 1 record at a time.

Arshad

 

[rrojo7229]

Hi,

No, this table is not a separate area, it is in the same area where all the tables are: "Schema Area".


Kind regards,

 

[tamhas]

Now you have another good reason for moving it to a separate area!

The for each approach is going to be the slowest, but have the least bi impact. Bundle in sets of 100 to speed it up.

 

[taqvia]

I just remember that once i dropped entire table without having the system to be compiled.

take the df of the table with crc option checked for r-code compatibility.
rename the existing table to x-table.
load the df it will have the same structure and r code compatibility. you can start your application with zero records in it.

Now you have ample time to get rid of x-table without any downtime as Thomas suggested have a tran block of 100 and delete it.

 

[TomBascom]

No, this table is not a separate area, it is in the same area where all the tables are: "Schema Area".

You really ought to change that... Some Surprising Benefits Of Using Storage Areas