public class ColumnPermutedStarTable extends WrapperStarTable
An int[] array, columnMap,
is used to keep track of
which columns in this table correspond to which columns in the base table;
the n'th column in this table corresponds to the
columnMap[n]'th column in the base table.
The columnMap array may contain duplicate entries, but all
its entries must be in the range
0..baseTable.getColumnCount()-1.
This table will have columnMap.length entries.
One implementation decision to be made is how to implement the
getRow methods on this table and on any RowSequence taken
out on it. It could either invoke getRow on the base
table/sequence, or invoke getCell multiple times.
The performance characteristics are likely to be different for the
two, though it's not obvious which will be best. In general,
if this table has the same number of columns (or more) than the
base table, then using getRow will be better, and if it
has many fewer, then getCell will be.
The user may choose between these behaviours using the
readRow parameter in one of the constructors.
If the other constructor (which doesn't have this argument) is used
then a choice will be made on the basis of the ratio of columns
in this table to the ratio of the columns in the base table
(the threshold is given by READROW_FRACTION).
| Modifier and Type | Field and Description |
|---|---|
static double |
READROW_FRACTION
When
readRow is not specified explicitly, this gives the
threshold for determining how getRow methods
are implemented. |
baseTable| Constructor and Description |
|---|
ColumnPermutedStarTable(StarTable baseTable,
int[] columnMap)
Constructs a new
ColumnPermutedStarTable
from a base StarTable and a columnMap array,
making its own decision about how to read rows. |
ColumnPermutedStarTable(StarTable baseTable,
int[] columnMap,
boolean readRow)
Constructs a new
ColumnPermutedStarTable
from a base StarTable and columnMap array,
with explicit instruction about how to read rows. |
| Modifier and Type | Method and Description |
|---|---|
static ColumnPermutedStarTable |
deleteColumns(StarTable baseTable,
int[] icols)
Returns a
ColumnPermutedStarTable with selected
columns deleted. |
java.lang.Object |
getCell(long irow,
int icol)
Returns the contents of a given table cell.
|
int |
getColumnCount()
Returns the number of columns in this table.
|
ColumnInfo |
getColumnInfo(int icol)
Returns the object describing the data in a given column.
|
int[] |
getColumnMap()
Returns the mapping used to define the permutation of the columns
of this table with respect to the base table.
|
java.lang.Object[] |
getRow(long irow)
Returns the contents of a given table row.
|
RowAccess |
getRowAccess()
Returns an object which can provide random access to this
table's data, if random access is implemented.
|
RowSequence |
getRowSequence()
Returns an object which can iterate over all the rows in the table
sequentially.
|
RowSplittable |
getRowSplittable()
Returns an object which can iterate over all the rows in the table,
but which may also be requested to split recursively
for potentially parallel processing.
|
void |
setColumnMap(int[] columnMap)
Sets the mapping used to define the permutation of the columns
of this table with respect to the base table.
|
checkedLongToInt, close, getBaseTable, getColumnAuxDataInfos, getName, getParameterByName, getParameters, getRowCount, getURL, isRandom, setName, setParameter, setURL, toStringpublic static final double READROW_FRACTION
readRow is not specified explicitly, this gives the
threshold for determining how getRow methods
are implemented.
If
getColumnCount() / getBaseTable().getColumnCount() > READROW_FRACTIONthen the base
getRow method is implemented, otherwise
multiple calls of getCell are used.public ColumnPermutedStarTable(StarTable baseTable, int[] columnMap)
ColumnPermutedStarTable
from a base StarTable and a columnMap array,
making its own decision about how to read rows.baseTable - the table on which this one is basedcolumnMap - array describing where each column of this table
comes from in baseTablepublic ColumnPermutedStarTable(StarTable baseTable, int[] columnMap, boolean readRow)
ColumnPermutedStarTable
from a base StarTable and columnMap array,
with explicit instruction about how to read rows.baseTable - the table on which this one is basedcolumnMap - array describing where each column of this table
comes from in baseTablereadRow - true if the readRow
implementations should
invoke readRow on the base table, false if they
should use readCell insteadpublic int[] getColumnMap()
public void setColumnMap(int[] columnMap)
columnMap - column permutation mappublic int getColumnCount()
StarTablegetColumnCount in interface StarTablegetColumnCount in class WrapperStarTablepublic ColumnInfo getColumnInfo(int icol)
StarTablegetColumnInfo in interface StarTablegetColumnInfo in class WrapperStarTableicol - the column for which header information is requiredicolpublic RowSequence getRowSequence() throws java.io.IOException
StarTablegetRowSequence in interface StarTablegetRowSequence in class WrapperStarTablejava.io.IOException - if there is an error providing accesspublic RowAccess getRowAccess() throws java.io.IOException
StarTablegetRowAccess in interface StarTablegetRowAccess in class WrapperStarTablejava.io.IOException - if there is an error setting up accesspublic RowSplittable getRowSplittable() throws java.io.IOException
StarTableThe return value must be non-null, and may provide splitting
arrangements specially appropriate for the implementation.
If this table 'wraps' an upstream table, it is usually best to
base the implementation on calls to the the upstream
getRowSplittable method, so that upstream policy
about how to divide up the table is respected.
However, implementations without special requirements may return
Tables.getDefaultRowSplittable(this).
getRowSplittable in interface StarTablegetRowSplittable in class WrapperStarTablejava.io.IOExceptionRowRunnerpublic java.lang.Object getCell(long irow,
int icol)
throws java.io.IOException
StarTablegetColumnInfo(icol).getContentClass().
This method is safe for concurrent use from multiple threads,
but in general it is recommended to use a
RowAccess instead.
getCell in interface StarTablegetCell in class WrapperStarTableirow - the index of the cell's rowicol - the index of the cell's columnjava.io.IOException - if there is an error reading the datapublic java.lang.Object[] getRow(long irow)
throws java.io.IOException
StarTablegetCell(irow,icol) for all the columns icol
in sequence.
This method is safe for concurrent use from multiple threads,
but in general it is recommended to use a
RowAccess instead.
getRow in interface StarTablegetRow in class WrapperStarTableirow - the index of the row to retrieveirowjava.io.IOException - if there is an error reading the datapublic static ColumnPermutedStarTable deleteColumns(StarTable baseTable, int[] icols)
ColumnPermutedStarTable with selected
columns deleted.
This utility method simply works out a permutation map which
corresponds to deletion of the columns indexed by the elements
of the icols array and passes it to the constructor.
A single column index may harmlessly be listed multiple times
in the icols array. It is not permitted for any
element of icols to be less than zero or
greater or equal than the number of columns in baseTable.baseTable - table whose columns are to be deletedicols - array with elements listing the column indices
which are to be removed