org.h2.index

Interface Index

All Superinterfaces:

DbObject, SchemaObject

All Known Subinterfaces:

MVIndex, SpatialIndex

All Known Implementing Classes:

BaseIndex, FunctionIndex, HashIndex, LinkedIndex, MetaIndex, MultiVersionIndex, MVDelegateIndex, MVPrimaryIndex, MVSecondaryIndex, MVSpatialIndex, NonUniqueHashIndex, PageBtreeIndex, PageDataIndex, PageDelegateIndex, PageIndex, RangeIndex, ScanIndex, SpatialTreeIndex, TreeIndex, ViewIndex

public interface Index
extends SchemaObject

An index. Indexes are used to speed up searching data.

Field Summary

Fields inherited from interface org.h2.engine.DbObject

AGGREGATE, COMMENT, CONSTANT, CONSTRAINT, FUNCTION_ALIAS, INDEX, RIGHT, ROLE, SCHEMA, SEQUENCE, SETTING, SYNONYM, TABLE_OR_VIEW, TRIGGER, USER, USER_DATATYPE

Method Summary

Methods

Modifier and Type	Method and Description
void	add(Session session, Row row) Add a row to the index.
boolean	canFindNext() Check if the index can get the next higher value.
boolean	canGetFirstOrLast() Check if the index can directly look up the lowest or highest value of a column.
boolean	canScan() Can this index iterate over all rows?
void	close(Session session) Close this index.
void	commit(int operation, Row row) Commit the operation for a row.
int	compareRows(SearchRow rowData, SearchRow compare) Compare two rows.
IndexLookupBatch	createLookupBatch(TableFilter[] filters, int filter) Creates new lookup batch.
Cursor	find(Session session, SearchRow first, SearchRow last) Find a row or a list of rows and create a cursor to iterate over the result.
Cursor	find(TableFilter filter, SearchRow first, SearchRow last) Find a row or a list of rows and create a cursor to iterate over the result.
Cursor	findFirstOrLast(Session session, boolean first) Find the first (or last) value of this index.
Cursor	findNext(Session session, SearchRow higherThan, SearchRow last) Find a row or a list of rows that is larger and create a cursor to iterate over the result.
int	getColumnIndex(Column col) Get the index of a column in the list of index columns
Column[]	getColumns() Get the indexed columns.
double	getCost(Session session, int[] masks, TableFilter[] filters, int filter, SortOrder sortOrder, java.util.HashSet<Column> allColumnsSet) Estimate the cost to search for rows given the search mask.
long	getDiskSpaceUsed() Get the used disk space for this index.
IndexColumn[]	getIndexColumns() Get the indexed columns as index columns (with ordering information).
IndexType	getIndexType() Get the index type.
java.lang.String	getPlanSQL() Get the message to show in a EXPLAIN statement.
Row	getRow(Session session, long key) Get the row with the given key.
long	getRowCount(Session session) Get the row count of this table, for the given session.
long	getRowCountApproximation() Get the approximated row count for this table.
Table	getTable() Get the table on which this index is based.
boolean	isFindUsingFullTableScan() Returns true if find() implementation performs scan over all index, false if find() performs the fast lookup.
boolean	isFirstColumn(Column column) Check if the given column is the first for this index
boolean	isRowIdIndex() Does this index support lookup by row id?
boolean	needRebuild() Check if the index needs to be rebuilt.
void	remove(Session session) Remove the index.
void	remove(Session session, Row row) Remove a row from the index.
void	setSortedInsertMode(boolean sortedInsertMode) Enable or disable the 'sorted insert' optimizations (rows are inserted in ascending or descending order) if applicable for this index implementation.
void	truncate(Session session) Remove all rows from the index.

Methods inherited from interface org.h2.schema.SchemaObject

getSchema, isHidden

Methods inherited from interface org.h2.engine.DbObject

checkRename, getChildren, getComment, getCreateSQL, getCreateSQLForCopy, getDatabase, getDropSQL, getId, getName, getSQL, getType, isTemporary, removeChildrenAndResources, rename, setComment, setTemporary

Method Detail

getPlanSQL

java.lang.String getPlanSQL()

Get the message to show in a EXPLAIN statement.

Returns:

the plan

close

void close(Session session)

Close this index.

Parameters:

session - the session used to write data

add

void add(Session session,
       Row row)

Add a row to the index.

Parameters:

session - the session to use

row - the row to add

remove

void remove(Session session,
          Row row)

Remove a row from the index.

Parameters:

session - the session

row - the row

isFindUsingFullTableScan

boolean isFindUsingFullTableScan()

Returns true if find() implementation performs scan over all index, false if find() performs the fast lookup.

Returns:

true if find() implementation performs scan over all index, false if find() performs the fast lookup

find

Cursor find(Session session,
          SearchRow first,
          SearchRow last)

Find a row or a list of rows and create a cursor to iterate over the result.

Parameters:

session - the session

first - the first row, or null for no limit

last - the last row, or null for no limit

Returns:

the cursor to iterate over the results

find

Cursor find(TableFilter filter,
          SearchRow first,
          SearchRow last)

Find a row or a list of rows and create a cursor to iterate over the result.

Parameters:

filter - the table filter (which possibly knows about additional conditions)

first - the first row, or null for no limit

last - the last row, or null for no limit

Returns:

the cursor to iterate over the results

getCost

double getCost(Session session,
             int[] masks,
             TableFilter[] filters,
             int filter,
             SortOrder sortOrder,
             java.util.HashSet<Column> allColumnsSet)

Estimate the cost to search for rows given the search mask. There is one element per column in the search mask. For possible search masks, see IndexCondition.

Parameters:

session - the session

masks - per-column comparison bit masks, null means 'always false', see constants in IndexCondition

filters - all joined table filters

filter - the current table filter index

sortOrder - the sort order

allColumnsSet - the set of all columns

Returns:

the estimated cost

remove

void remove(Session session)

Remove the index.

Parameters:

session - the session

truncate

void truncate(Session session)

Remove all rows from the index.

Parameters:

session - the session

canGetFirstOrLast

boolean canGetFirstOrLast()

Check if the index can directly look up the lowest or highest value of a column.

Returns:

true if it can

canFindNext

boolean canFindNext()

Check if the index can get the next higher value.

Returns:

true if it can

findNext

Cursor findNext(Session session,
              SearchRow higherThan,
              SearchRow last)

Find a row or a list of rows that is larger and create a cursor to iterate over the result.

Parameters:

session - the session

higherThan - the lower limit (excluding)

last - the last row, or null for no limit

Returns:

the cursor

findFirstOrLast

Cursor findFirstOrLast(Session session,
                     boolean first)

Find the first (or last) value of this index. The cursor returned is positioned on the correct row, or on null if no row has been found.

Parameters:

session - the session

first - true if the first (lowest for ascending indexes) or last value should be returned

Returns:

a cursor (never null)

needRebuild

boolean needRebuild()

Check if the index needs to be rebuilt. This method is called after opening an index.

Returns:

true if a rebuild is required.

getRowCount

long getRowCount(Session session)

Get the row count of this table, for the given session.

Parameters:

session - the session

Returns:

the row count

getRowCountApproximation

long getRowCountApproximation()

Get the approximated row count for this table.

Returns:

the approximated row count

getDiskSpaceUsed

long getDiskSpaceUsed()

Get the used disk space for this index.

Returns:

the estimated number of bytes

compareRows

int compareRows(SearchRow rowData,
              SearchRow compare)

Compare two rows.

Parameters:

rowData - the first row

compare - the second row

Returns:

0 if both rows are equal, -1 if the first row is smaller, otherwise 1

getColumnIndex

int getColumnIndex(Column col)

Get the index of a column in the list of index columns

Parameters:

col - the column

Returns:

the index (0 meaning first column)

isFirstColumn

boolean isFirstColumn(Column column)

Check if the given column is the first for this index

Parameters:

column - the column

Returns:

true if the given columns is the first

getIndexColumns

IndexColumn[] getIndexColumns()

Get the indexed columns as index columns (with ordering information).

Returns:

the index columns

getColumns

Column[] getColumns()

Get the indexed columns.

Returns:

the columns

getIndexType

IndexType getIndexType()

Get the index type.

Returns:

the index type

getTable

Table getTable()

Get the table on which this index is based.

Returns:

the table

commit

void commit(int operation,
          Row row)

Commit the operation for a row. This is only important for multi-version indexes. The method is only called if multi-version is enabled.

Parameters:

operation - the operation type

row - the row

getRow

Row getRow(Session session,
         long key)

Get the row with the given key.

Parameters:

session - the session

key - the unique key

Returns:

the row

isRowIdIndex

boolean isRowIdIndex()

Does this index support lookup by row id?

Returns:

true if it does

canScan

boolean canScan()

Can this index iterate over all rows?

Returns:

true if it can

setSortedInsertMode

void setSortedInsertMode(boolean sortedInsertMode)

Enable or disable the 'sorted insert' optimizations (rows are inserted in ascending or descending order) if applicable for this index implementation.

Parameters:

sortedInsertMode - the new value

createLookupBatch

IndexLookupBatch createLookupBatch(TableFilter[] filters,
                                 int filter)

Creates new lookup batch. Note that returned IndexLookupBatch instance can be used multiple times.

Parameters:

filters - the table filters

filter - the filter index (0, 1,...)

Returns:

created batch or null if batched lookup is not supported by this index.