----- Original Message -----
> From: "Rik Wasmus" <rik@stripped>
>
> Depends on the data and usage, but probably I'd go for a index(a,b) &
> index(b,a) if reads heavily outnumber writes. As index(a) is covered
> by index(a,b), and index(b) by index(b,a), we don't need to add those,
> which saves time on modifications.
I'm trying to think of a scenario where index(a) would be beneficial in the presence of
index(a,b). If both are available, and all else being equal, it's likely that the parser
will pick the simplest index; but I can't see it having a major impact.
Any full prefix of a combined index may be used; so afaik a separate index on any full
prefix is a waste of diskspace and cycles.
The net conclusion, Neil, is that you actually have to know what you're doing :-) Take the
time to read the online documentation on mysql.com, it's pretty good.
--
Bier met grenadyn
Is als mosterd by den wyn
Sy die't drinkt, is eene kwezel
Hy die't drinkt, is ras een ezel
