SQL*Loader suggestion for huge records [message #140274] |
Mon, 03 October 2005 06:29 |
himang
Messages: 282 Registered: March 2005 Location: Bangalore
|
Senior Member |
|
|
I have a flat file which has around 1 million records. There are no delimiter in the records and whole line has to be
considered as 1 column. There are around 8,00,000 records starting with "Warning" and rest starting with "Start" or "End".
I am loading all the records in 1 temporary table and I need only count of records starting with "Warning" and for other
I have to do some calculation.
what other option can I use?
1. Should I load the data in seperate tables for seperate parameters (1 for Warning ones and other for "Start/End")? Will
it have any performance problem while loading?
2. Or should I load in single table and then load in main tables after doing the calculation? This will result in
degrade of performance while doing calculation?
Any suggestions?
File looks like this:-
START: read routing_data (NULL) (03:01:00)
END: read routing_data (NULL) (03:01:00): 0.000 sec
Warning: deleting mfg order 515763300_1-001 which is no longer pegged.
Warning: deleting mfg order 515885400_1-001 which is no longer pegged.
Warning: deleting mfg order 518833100_1-001 which is no longer pegged.
Warning: deleting mfg order 519245700_1-001 which is no longer pegged.
Regards
Himanshu
|
|
|
|
|
|