database question

[gsh]

I'm creating a *huge* table in mysql, at least 50 Gig. I have a C program reading a file from else where, picking out the info I want/need, then stuffing it into mysql with lots of insert into table (ra,dec,mag....) values () where ra and dec are indexed.

I noticed the amount of time it takes to do the inserts increases.

Does anyone have any advice on if it is faster to create my table with indicies, and stuff away, or should I create my table without indecies,
slurp in the 50-75 gig of data, then ALTER TABLE ADD INDEX?

Comments

[gsh]

Mysql has a bulk data loader thingy, but I'm only trying to load a subset of the data.
I don't mind it running all weekend.

I'll drop the table, and try again.

[dckermit.livejournal.com]

My experience, also with other applications, is that it's far faster to have the extraction program create a flat file from the "elsewhere" data, then allow the database application to load that flat file in bulk. Sometimes days faster.

[gsh]

Yea, I'm doing "creative lazyness" and hoping to avoid that. I started a shell script
to load the data. If its all loaded by monday 9 am i'll call it victory.