
28 Sep
2010
28 Sep
'10
10:58 a.m.
On 28/09/10 11:49, Stewart, Robert wrote:
Francisco José Tapia wrote:
Is there any interest in a library implementing counter trees, which permit us to access to the elements by the position, as in the same way than a vector. The insertion, deletion and access to elements are operations O(log N).
How does this differ from Boost.MultiIndex which can provide random access plus ordered, key-based indices in a single container?
AFAIK, in that case Boost.MultiIndex stores the data and an index separately as two distinct structures. Counter trees provide both features in the same data structure; which might be a good idea or not.