Home > database >  Large amount of data, please insert (or merge into), what methods of speed?
Large amount of data, please insert (or merge into), what methods of speed?

Time:09-25

The great spirit is good, I now have a problem, want to oracle in hundreds of millions of data need to index points now area very slow, I still have what way to improve the speed? At present all partitions in a table space, the speed of the first after have 2 w to 10 million per second per second only 23000

This is my create table statements, are now in a table space


I use the merge into the method of inserting (almost) and insert into efficiency
Mergeinto statement below


Vc_bf_foid and vc_df_foid index index was made in the field is a must, because in addition to the insert need to do the update operation in these two fields,
The great spirit returns busy, this problem bothering me for a long time, various tests are slow

CodePudding user response:


Fill a table statements diagram constructed this is I now test table

CodePudding user response:

Try to partition.

CodePudding user response:

refer to the second floor wmxcn2000 response:
try to partition,
great god to dwell on it for me? Transferred to partition mean?

CodePudding user response:

The great god for help!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

CodePudding user response:

See if the table space is full, have been automatic extension resulting in slow, UNDO has eyes to see

CodePudding user response:

Million level data insert, I used to test for many times, the insert billions of data can only be used parallel to ensure inventory rapidly in a certain period of time

CodePudding user response:

Bulk insert in the middle table, and then from the middle to your formal list table synchronization
  • Related