BTreeIndex class. It implements a B+ Tree index on a single attribute of a relation. This index supports only one scan at a time. More...

#include <btree.h>

Public Member Functions
	BTreeIndex (const std::string &relationName, std::string &outIndexName, BufferManager *bufMgrIn, const int attrByteOffset, const Datatype attrType)
	~BTreeIndex ()
const void	insertEntry (const void *key, const RecordId rid)
const void	startScan (const void lowVal, const Operator lowOp, const void highVal, const Operator highOp)
const void	scanNext (RecordId &outRid)
const void	endScan ()

Detailed Description

BTreeIndex class. It implements a B+ Tree index on a single attribute of a relation. This index supports only one scan at a time.

Definition at line 293 of file btree.h.

Constructor & Destructor Documentation

wiscdb::BTreeIndex::BTreeIndex	(	const std::string &	relationName,
		std::string &	outIndexName,
		BufferManager *	bufMgrIn,
		const int	attrByteOffset,
		const Datatype	attrType
	)

BTreeIndex Constructor. Check to see if the corresponding index file exists. If so, open the file. If not, create it and insert entries for every tuple in the base relation using FileScanner class.

Parameters:

relationName	Name of file.
outIndexName	Return the name of index file.
bufMgrIn	Buffer Manager Instance
attrByteOffset	Offset of attribute, over which index is to be built, in the record
attrType	Datatype of attribute over which index is built

Exceptions:

BadIndexInfoException If the index file already exists for the corresponding attribute, but values in metapage(relationName, attribute byte offset, attribute type etc.) do not match with values received through constructor parameters.

Definition at line 29 of file btree.cpp.

wiscdb::BTreeIndex::~BTreeIndex ( )

BTreeIndex Destructor. End any initialized scan, flush index file, after unpinning any pinned pages, from the buffer manager and delete file instance thereby closing the index file. Destructor should not throw any exceptions. All exceptions should be caught in here itself.

Definition at line 43 of file btree.cpp.

Member Function Documentation

const void wiscdb::BTreeIndex::endScan ( )

Terminate the current scan. Unpin any pinned pages. Reset scan specific variables.

Exceptions:

ScanNotInitializedException If no scan has been initialized.

Definition at line 81 of file btree.cpp.

const void wiscdb::BTreeIndex::insertEntry	(	const void *	key,
		const RecordId	rid
	)

Insert a new entry using the pair <value,rid>. Start from root to recursively find out the leaf to insert the entry in. The insertion may cause splitting of leaf node. This splitting will require addition of new leaf page number entry into the parent non-leaf, which may in-turn get split. This may continue all the way upto the root causing the root to get split. If root gets split, metapage needs to be changed accordingly. Make sure to unpin pages as soon as you can.

Parameters:

key	Key to insert, pointer to integer/double/char string
rid	Record ID of a record whose entry is getting inserted into the index.

Definition at line 51 of file btree.cpp.

const void wiscdb::BTreeIndex::scanNext ( RecordId & outRid )

Fetch the record id of the next index entry that matches the scan. Return the next record from current page being scanned. If current page has been scanned to its entirety, move on to the right sibling of current page, if any exists, to start scanning that page. Make sure to unpin any pages that are no longer required.

Parameters:

outRid RecordId of next record found that satisfies the scan criteria returned in this

Exceptions:

ScanNotInitializedException	If no scan has been initialized.
IndexScanCompletedException	If no more records, satisfying the scan criteria, are left to be scanned.

Definition at line 72 of file btree.cpp.

const void wiscdb::BTreeIndex::startScan	(	const void *	lowVal,
		const Operator	lowOp,
		const void *	highVal,
		const Operator	highOp
	)

Begin a filtered scan of the index. For instance, if the method is called using ("a",GT,"d",LTE) then we should seek all entries with a value greater than "a" and less than or equal to "d". If another scan is already executing, that needs to be ended here. Set up all the variables for scan. Start from root to find out the leaf page that contains the first RecordID that satisfies the scan parameters. Keep that page pinned in the buffer pool.

Parameters:

lowVal	Low value of range, pointer to integer / double / char string
lowOp	Low operator (GT/GTE)
highVal	High value of range, pointer to integer / double / char string
highOp	High operator (LT/LTE)

Exceptions:

BadOpcodesException	If lowOp and highOp do not contain one of their their expected values
BadScanrangeException	If lowVal > highval
NoSuchKeyFoundException	If there is no key in the B+ tree that satisfies the scan criteria.

Definition at line 60 of file btree.cpp.

The documentation for this class was generated from the following files:

Public Member Functions

Detailed Description

Constructor & Destructor Documentation

Member Function Documentation