| Version 1.9 Build 1367
|
|
Next: Current capabilities and design
Up: MeasurementSet access in AIPS++
Previous: Introduction
MeasurementSets are AIPS++ Tables and their existing access
classes are built closely on the Table infrastructure. This
assumption is implicit in the requirements listed in this section.
The C++ classes used for MS access need to meet the following
requirements, or provide the following services:
- 1.
- Basic whole MS operations: MS creation, deletion and
related MS-level operations.
- 2.
- Column and row access: Access to MS MAIN and sub-table
rows and columns.
- 3.
- Sorting and selection: The ability to support bulk sorting
and selection operations. MS selections are expected to be expressed
in a close variant of the Table Query Language (TaQL), defined
elsewhere.
- 4.
- MAIN MS iteration: Sequential iteration through the MAIN
MS table on arbitrary iteration indices, with optional retrieval of
associated MS sub-table data matching the current MAIN table iteration
interval. Indexed access to the MAIN table, on limited key indices, is
a secondary requirement. MAIN table iteration includes simultaneous
or sequential iteration through multiple MS considered as a
set. Iteration is also required within MS MAIN rows, in frequency,
velocity or polarization.
- 5.
- MS sub-table iteration: Independent sequential iteration
through individual MS sub-tables, on arbitrary iteration
indices. Keyed access to MS sub-tables in read-write mode, on
arbitrary key indices. Iteration is also required within rows, in
frequency, velocity or polarization, as appropriate for individual
sub-tables.
- 6.
- MS derived quantities: Computation or retrieval of derived
MS quantities, which require data access across the MS as a
whole. Examples include elevation computation, assembling higher-level
coordinate information or retrieval of Doppler shift parameters.
- 7.
- MS data ranges: Determination of the data ranges in a
given MS, for any MAIN or sub-table columns or derived quantities.
- 8.
- MS data objects: Provision of C++ classes to model a data
cube containing part or all of the current iteration or selection,
which can be operated upon by other agent classes (e.g. calibration),
and passed between AIPS++ classes and between C++ and Glish.
- 9.
- Connection to Glish: Versatile access to the C++ MS access
classes from Glish.
- 10.
- Scratch column management: Creation and deletion of
scratch columns which may be added to, or removed from the MS during
reduction.
- 11.
- Parallel I/O support: Full integration of parallel I/O
capabilities within the standard serial MS access classes, and the
ability to optionally enable the parallel I/O capabilities.
- 12.
- Table infrastructure use: Full re-use of all Table system infrastructure and design philosophy wherever possible
in the MS access classes.
- 13.
- Efficiency: I/O is a critical part of astronomical data
reduction performance, and efficiency is a key requirement for MS
access. This includes integrated I/O profiling, in both the serial
and parallel case.
Next: Current capabilities and design
Up: MeasurementSet access in AIPS++
Previous: Introduction
  Contents
Please send questions or comments about AIPS++ to aips2-request@nrao.edu.
Copyright © 1995-2000 Associated Universities Inc.,
Washington, D.C.
Return to AIPS++ Home Page
2006-03-28