myisam issues on indexing

2001-06-22 Thread Robin Senior
Hi, While trying to index a 77million row table (I know, I'll shut up), I started to fiddle with the mysql variables to get it done in under 3 weeks. Finally, I happened upon the myisam_max_sort_file_size and myisam_max_extra_sort_file_size variables. Messing with these allowed the indexing to go

Re: indexing = NP complete??

2001-06-19 Thread Robin Senior
On Fri, 15 Jun 2001, Jeremy Zawodny wrote: > On Fri, Jun 15, 2001 at 12:53:04PM -0600, Chris Bolt wrote: > > > Is this machine basically dedicated to MySQL? If so, I'd try bumping > > > up the record_buffer to 16 or 32MB and see what that does. I've got it > > > set at 32MB on a server with 1GB o

Re: indexing = NP complete??

2001-06-14 Thread Robin Senior
On Thu, 14 Jun 2001, Jeremy Zawodny wrote: > I have a few comments and questions: > > (1) Is this a unique index? If not, consider doing a partial index > on the column. Index the first 32 characters or so. I'll save a > lot of disk I/O, memory, and CPU time but still yeild pretty

indexing = NP complete??

2001-06-14 Thread Robin Senior
Just kidding! But it seems that way... Indexing a 7 million row VARCHAR(255) table took ~12 minutes on my Athlon 1.33Ghz w/512MB. Indexing a 77 million row VARCHAR(255) table has taken 9 DAYS, and still hasn't finished. Obviously this isn't a linearly scaling operation, but it is seeming dang ne

indexing speed

2001-06-12 Thread Robin Senior
Hi, I'm running MySQLd-NT on an Athlon 1.33GHz w/512MB RAM, under Win2000. I'm doing a fulltext index on a varchar(255) column in one of my tables. Admittedly, the table has ~77million rows, but the indexing has been going for almost a *week* now. What can I do to optimize MySQL for fulltext inde