How to HPL to pipe to HPL?
Posted in 1999
Topics: Migration, Import/Export & Data Conversion
Hi, We need to copy/move 24 million x 200 byte row sized tables around on a regular basis. The two main options we use right now are: 1) the blunt INSERT INTO table_copy SELECT required_columns FROM original_table WHERE certain_filters, 2) the almost just as blunt, HPL unload of the SELECT and and HPL load into the copy_table. The 2nd option usually turns out to be the quicker. We would like to figure out how we can have the HPL unload feed the HPL load directly through a pipe of sorts so we do not have to wait upon completion of the unload before starting the load. Has anybody got an idea how to best do this, or if this is at all possible? This could really make a killer data movement solution for our situation.... Thanks. Arnoud.
Arnoud, Could you post what exactly you are trying to accomplish by these periodic reloads? It sounds to me that you are doing something the hard way. BTW You may also find my dbcopy.ec utility to be as fast as or faster than HPL to disk the reload HPL from disk. It averages 3X as fast as INSERT INTO....SELECT FROM.... with the -F feature enabled! But still, there may be an easier way. I just do not know what you are hoping to accomplish. Art S. Kagel Arnoud Otte wrote: > > Hi, > > We need to copy/move 24 million x 200 byte row sized tables around on a > regular basis. The two main options we use right now are: > > 1) the blunt INSERT INTO table_copy SELECT required_columns FROM > original_table WHERE certain_filters, > 2) the almost just as blunt, HPL unload of the SELECT and and HPL load into > the copy_table. > > The 2nd option usually turns out to be the quicker. We would like to figure > out how we can have the HPL unload feed the HPL load directly through a pipe > of sorts so we do not have to wait upon completion of the unload before > starting the load. > > Has anybody got an idea how to best do this, or if this is at all possible? > This could really make a killer data movement solution for our situation.... > > Thanks. > > Arnoud.