- Amp
- Phase
- Real
- Imag
- Data
- nFuncType
- Observed
- Corrected
- Model
- Ratio
- Residual
- ObsResidual
- ObsFloat
- nDataType
construct from an MS, the MS will supply the range of the various
parameters that can be selected on.
Copy constructor, this will initialize the MS with other's MS
Assignment, this will initialize the MS with other's MS
Change or Set the MS this MSSelector refers to.
initialize the selection by specifying, optionally,
the DATA_DESC_IDs.
If you don't specify the dataDescIds and the data shape is constant
all data is selected, if the shape does change, only the first
dataDescId is selected. If you specify a number of dataDescIds
and they all have the same shape, they are all selected, otherwise
only the first is selected. The function returns false if
the selection was limited due to changing data shape.
Use the reset argument to return to the completely unselected ms.
As above without the data desc id argument
Return the data desc IDs selected
Set the mapping from input channels in the DATA column to
output channels. nChan is the number of output channels,
start is the first channel to use, width specifies how wide a
block of channels to average, increment specifies the start of
the next block relative to the start of the current block.
Note: averaging uncalibrated data should be avoided (no bandpass applied)
Specify the output polarization.
Missing input polarizations are assumed to be zero.
This selection/conversion assumes that parallactic angle rotation
is taken care of elsewhere (i.e., results may only be correct for
CORRECTED_DATA and MODEL_DATA conversions, not for the observed DATA)
Select the MS based on the selections present in the input record.
The format of this record is the same as that returned by range.
Not all possible items can be selected on, some are quietly ignored.
Correct for one-based indexing if oneBased is True.
Select the MS based on the TaQL selection string
Return the data for the items requested, all returned values
will be arrays, the last dimension of these is the table row number.
The data arrays are normally 3D with axes: polarization, frequency, row.
If ifrAxis is set to True, the data arrays returned will be 4D, with
the data being split out along an extra interferometer axis, the
axes will be: polarization, frequency, interferometer and time.
Missing interferometers will be marked flagged.
The order of the interferometers is that specified by the last
select call.
Add a (flagged) gap in the data at every antenna1 change if ifrAxisGap>0.
Use inc > 1 to return data from every inc'th row.
Use average=True to vector average the data along the row or time axis
taking the weights column into account (use selectChannel to average
channels together as well). Note that different interferometers will be
averaged together if ifrAxis is False.
Correct for one-based indexing if oneBased is True.
Put the data for the items provided. Note that only fields corresponding
to actual table columns can be put (i.e., no AMPLITUDEs, IFR_NUMBERs etc)
The data will need to have the correct shape for the column and a last
dimension matching the number of selected rows (or last two dimensions
matching times and interferometers, for data retrieved with ifraxis=T)
Channel selection is supported, but the width parameter has to be 1.
Set up an iterator, iterating over the specified columns, with
optional time interval and maximum number of rows to return at once
(the default of zero returns all rows). To keep MSIter from adding
the default sort columns, specify addDefaultSortColumns=False
Step the iterator, sets the selection to the current table iteration.
Returns false if there is no more data
and sets the selection back to the state before iteration started.
(Re)Set the iterator to the first iteration, call this after iterInit.
/ End the iteration (before reaching the last iteration)
and set the selection back to the state before iteration started.
Number of rows in selected table
Return the selected table
Return the selection status of the table
average and convert float data
average and convert data, with row Slicer
average and convert float data, with row Slicer
"average" flag, at present all output which has a flagged input is flagged
"average" flag, at present all output which has a flagged input is flagged,
with row Slicer
"unaverage" flag, distribute the flags back to the channels that went
into the average
get the weight, set sigma=True when retrieving sigma's
make the data slicer, pass in the first and the number of correlations
to select
reorder from 2d to 1d (removing ifr axis)
reorder from 2d to 1d (removing ifr axis)
time average the input data, return new flags
check if the data description selection has been done & do default
selection if not. Return False if the selection fails.
Review Status
Prerequisite
Etymology
MSSelector's main function is selection of data from a MeasurementSet
Synopsis
This class is used to select and retrieve data from a MeasurementSet.
It allows selections on e.g., time, field, spectral window (all row based),
but also on channel and polarization (within a row). It can optionally
do polarization conversion, spectral averaging and time averaging on the
data retrieved and allows modified data to be written back to the Table.
This class also provides the DO interface to the MS Iterator.
The ms DO uses this class to allow these operations to be done from glish.
MSSelector msSelector(myMS);
// select data desc Id 1
msSelector.initSelection(1);
Vector<String> items(3);
// fill in some fields
items(0)="field_id";
items(1)="time";
items(2)="num_chan";
// get the range of values for the items specified
MSRange msRange(msSelector.selectedTable(),msSelector.spectralWindow());
Record range=msRange.range(items);
//.. change the ranges as needed
// now select with the new range
msSelector.select(range);
Int nchan=10, start=3, width=1, incr=2;
msSelector.selectChannel(nchan,start,width,incr)
// get out some data
Vector<String> dataItems(3);
dataItems(0)="data";
dataItems(1)="antenna1";
dataItems(2)="antenna2";
Record dataRec=msSelector.getData(items);
Motivation
Selection from an MS is needed in various places. It makes sense to
provide a uniform interface for MS selection.
Thrown Exceptions
To Do
Member Description
MSSelector()
explicit MSSelector(MeasurementSet& ms)
MSSelector(const MSSelector& other)
MSSelector& operator=(const MSSelector& other)
~MSSelector()
void setMS(MeasurementSet& ms)
Bool initSelection(const Vector<Int>& dataDescIds, Bool reset=False)
Bool initSelection(Bool reset=False)
Vector<Int> dataDescId() const
Bool selectChannel(Int nChan, Int start, Int width, Int incr)
Bool selectPolarization(const Vector<String>& wantedPol)
Bool select(const Record& items, Bool oneBased=False)
Bool select(const String& msSelect)
Record getData(const Vector<String>& items, Bool ifrAxis, Int ifrAxisGap=0, Int inc=1, Bool average=False, Bool oneBased=False)
Bool putData(const Record& items)
Bool iterInit(const Vector<String>& columns, Double interval, Int maxRows=0, Bool addDefaultSortColumns=True)
Bool iterNext()
Bool iterOrigin()
Bool iterEnd()
Int nrow() const
Table selectedTable() const
Bool selected() const
void getAveragedData(Array<Complex>& avData, const Array<Bool>& flag, const ROArrayColumn<Complex>& col) const
average and convert data
void getAveragedData(Array<Float>& avData, const Array<Bool>& flag, const ROArrayColumn<Float>& col) const
void getAveragedData(Array<Complex>& avData, const Array<Bool>& flag, const ROArrayColumn<Complex>& col, const ROColumn & rowSlicer) const
void getAveragedData(Array<Float>& avData, const Array<Bool>& flag, const ROArrayColumn<Float>& col, const Slicer & rowSlicer) const
Array<Bool> getAveragedFlag(Array<Bool>& avFlag, const ROArrayColumn<Bool>& col) const
Array<Bool> getAveragedFlag(Array<Bool>& avFlag, const ROArrayColumn<Bool>& col, const Slicer& rowSlicer) const
void putAveragedFlag(const Array<Bool>& avFlag, ArrayColumn<Bool>& col)
Array<Float> getWeight(const ROArrayColumn<Float>& wtCol, Bool sigma=False) const
void makeSlicer(Int start, Int nCorr) const
void reorderFlagRow(Array<Bool>& flagRow)
void reorderWeight(Array<Float>& weight)
void timeAverage(Array<Bool>& dataFlags, Array<Complex>& data, const Array<Bool>& flags, const Array<Float>& weights)
Bool checkSelection()
enum
The function types
enum
The data types