Re: [boost] boost::time_series

From: Eric Niebler (eric_at_[hidden])
Date: 2007-07-16 13:26:58

Tom Brinkman wrote:
> I wonder how efficient this method to populate the "time_series" containers is?
> 
>     make_ordered_inserter(s)
> 
>         (8, 2)
> 
>         (6, 4)
> 
>         (4, 6)
> 
>         (2, 8)
> 
>     .commit();
> 
> What is the preferred method to quickly populate a large number of
> data points, say in the thousands?

This is the preferred way to do it. It is algorithmically efficient 
because you are inserting sorted data into a series, so it doesn't need 
to do lookups. What is likely going on under the hood is a vector is 
being filled. This certainly has some overhead, and I could consider 
adding a reserve() customization point to the OrderedInserter concept.

> What is the preferred method to bind a structure's member variable to
> the ordered insertion list?
> 
> Example:
> 
> struct Foo
> {
>     boost::date_time::date date;
>     double high;
>     double low;
> };
> 
> Foo foo[10000];
> 
> How would I populate and bind "foo::high" into a sparse_series?

Not 100% sure I understand the question. Is "foo::high" supposed to be 
the value or the offset in the sparse series?

-- 
Eric Niebler
Boost Consulting
www.boost-consulting.com
The Astoria Seminar ==> http://www.astoriaseminar.com