Global Positioning System Reference
In-Depth Information
AML uses infi nite vector of integers for representing array data. Special
characteristics of the model are bit patterns. The core operators of AML
are defi ned with the help of bit patterns. AML has three core operators:
subsample , merge and apply . While the subsample operator eliminates cells
in an array, the merge operator combines two arrays and the apply operator
applies a given function to an array.
AQL
The third array model presented in this chapter is AQL , which relies on
Nested Relational Calculus for Arrays (NRCA) as extension of the Nested
Relational Calculus (NRC) (Libkin et al. 1996; Machlin 2007). The objective of
AQL was to support the application domains of the NetCDF data exchange
format and particularly scientifi c array data motivated by the fact to be able
to algorithmically generate and manipulate array data.
NRC contains objects like products and sets. The value set can be
generalized into an uninterpreted base type (i.e., black box). NRCA adds to
NRC natural numbers, constants, basic arithmetics, an index set generator,
and a summation construct. By this, arrays can be algorithmically generated
and manipulated.
NRCA introduces four operators: subscripting for returning the content
of a cell and length of a dimension, which both operate on arrays, and array
tabulation and indexing , which contribute to the generation of arrays.
RAM
As the last array model presented here, the Random-access memory (RAM)
was designed as extension to the neo-relational database management
system MonetDB (Cornacchia et al. 2008). Contrarily to AQL, this array
model is completely separated from the relational query formalism. Both
the motivation and the corresponding application area lay in multimedia
analysis (van Ballegooij 2004; van Ballegooij et al. 2005; Cornacchia et al.
2004). However, the model was meant for OLAP only and does not claim to
be suitable for geo services; SciQL is a successor which aims at incorporating
such application domains.
RAM offers a generic array construction operator similar to the array
tabulation of AQL. RAM introduces the concept of aligned arrays which are
defi ned as arrays with identical shape representing related data. Other six
operators are: const , grid , map , apply , choice and aggregate for fi lling a new
array with a constant value, applying a given function to aligned elements
in a set of arrays, applying a given array to aligned elements in a set of
index-arrays, choosing cell values from two input arrays and applying the
aggregation function along the fi rst given axes of an array.
Search WWH ::




Custom Search