Re: Slow ALTER TABLE on 70M row InnoDB table

2010-08-18 Thread Xn Nooby
>From what I have read, ALTER TABLE to add an index causes the entire table to be duplicated, so wouldn't my ALTER TABLE command be duplicating the work done by the SELECT command? On Wed, Aug 18, 2010 at 4:50 PM, mos wrote: > At 02:52 PM 8/18/2010, Xn Nooby wrote: >> >

Re: Slow ALTER TABLE on 70M row InnoDB table

2010-08-18 Thread Xn Nooby
; At 02:52 PM 8/18/2010, Xn Nooby wrote: >> >> Below is a generic version of the code I am trying.  It does copy the >> rows very quickly, but I will have to test to see how quickly the >> indices are built.  Is the below code what you were suggesting?  I had >> a littl

Re: Slow ALTER TABLE on 70M row InnoDB table

2010-08-18 Thread Xn Nooby
Hi Mike, my comments are below: On Wed, Aug 18, 2010 at 4:50 PM, mos wrote: > At 02:52 PM 8/18/2010, Xn Nooby wrote: >> >> Below is a generic version of the code I am trying.  It does copy the >> rows very quickly, but I will have to test to see how quickly the >>

Re: Slow ALTER TABLE on 70M row InnoDB table

2010-08-18 Thread Xn Nooby
Below is a generic version of the code I am trying. It does copy the rows very quickly, but I will have to test to see how quickly the indices are built. Is the below code what you were suggesting? I had a little trouble dropping and later adding the primary index, but I think I got it figured o

Slow ALTER TABLE on 70M row InnoDB table

2010-08-18 Thread Xn Nooby
I have been trying to speed up an ALTER TABLE command that adds a column to a large InnoDB table of about 80M rows. I have found and tried many different methods, but they are all slow.I have tried both optimizing the ALTER TABLE command, and dumping and loading the table (in both SQL and CSV for