[BioSQL-l] TAXON,TAXON_NAME, was Re: Description

[Hilmar Lapp]

You could theoretically save some time by not updating the values  
that wouldn't change (e.g., all left siblings and their descendants  
unchanged).

But on average, assuming that the new node is most likely a leaf node  
and that the tree is balanced, you would still have to update about  
half of all nodes, namely all nodes "to the right" (i.e., all  
siblings and their descendants to the right, all ancestors and their  
siblings to the right and the descendants of those siblings).

	-hilmar

On Sep 12, 2007, at 8:24 PM, Paul Davis wrote:

> I was more wondering if there was an efficient way to recompute that
> information. As you seem to be confirming, I was faily certain that to
> update those values would require recalculating all values.
>
> Paul

-- 
===========================================================
: Hilmar Lapp  -:-  Durham, NC  -:-  hlapp at gmx dot net :
===========================================================