Lopez David E-r9374c wrote:
matt
1) inserts using this format is much faster: INSERT INTO table (col1, col2) VALUES (val1,val2), (val3,val4) is much faster then single row insert. My experience is 2.5 hrs vs.. 36 hrs.
2) The PACK_KEYS=1 may be hurting you. I've never used it.
3) There may be a cache somewhere that's to small. You'll have to do some digging in this area.
4) dup key ignore - what does that mean exactly?
5) what is your OS & rev, mysql rev.
Please post any suggestions that you find valuable so we can all learn..
david
If I understand it, pack_keys should help if your disk IO limited vs cpu limited, cpu is ususally near idle.
I increased the cache sizes, helped a little, but not much
delay_key_write=ALL key_buffer_size=1000M read_buffer_size=512M record_buffer=512M
What would the syntax for that type of insert be?
I have a table with 30,000 records, I need to insert them into the main table with millions of records, I thought that insert into table select * from 2nd table would be the fastest way.
insert ignore will not insert a record, if it violates the primary key, I do this to keep duplicate records out of the system
windows 2003 and 2000 servers, mysql-nt 4.0.16 I tried the newer versions, but found bugs on all of them, I submitted them to the bug system.
I believe my bottleneck is reading the data to ensure the primary key is not violated, I see lots of read IO, but little write IO
Matt
-- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED]
