HashIndexScanner Class Reference

#include <hashindexscanner.h>

Collaboration diagram for HashIndexScanner:

Public Member Functions
	HashIndexScanner (HashIndexFile *index)
	Constructor for scanning all entries of the Index.
	HashIndexScanner (HashIndexFile *index, Key *key_val)
	Constructor for equality scan, this scan is over all K* entries with a matching key value to the passed key_val.
	~HashIndexScanner ()
	Destructor.
RecordId	getNext ()
	Returns RecordId of the next matching K* in the scan.
RecordId	getNext (Key *key)
	Returns RecordId of the next matching K* and sets the data part of the passed Key to the Key value of the next matching K* entry in the scan.

Private Attributes
HashIndexFile *	index
bool	full_scan
Key *	search_key
BufferManager *	buf_mgr
PageId	cur_pid
HashBucketPage *	cur_page
HashBucketPageScanner *	bucket_scanner
PageId	cur_dir_pid
bool	scan_complete

Detailed Description

HashIndexScanner: Scanner for Hash indicies. Different constructors are used to initiate different types of scans over indicies. getNext gets the the next matching tuple based on the type of scan

Constructor & Destructor Documentation

HashIndexScanner() [1/2]

HashIndexScanner::HashIndexScanner

(

HashIndexFile *

index

)

Constructor for scanning all entries of the Index.

Precondition

Valid HashIndexFile* is provided as input.

Postcondition

IndexFileScanner object is constructed and initialized to do a scan of the entire index. cur_page, the first page of index K* entries is pinned if file is not empty.

Parameters

index

IndexFile object to be scanned.

HashIndexScanner() [2/2]

HashIndexScanner::HashIndexScanner	(	HashIndexFile *	index,
		Key *	key_val
	)

Constructor for equality scan, this scan is over all K* entries with a matching key value to the passed key_val.

Precondition

Valid IndexFile and Key are provided as input.

Postcondition

HashIndexScanner object is constructed and initialized to do a scan of K* index entries with a matching key value. cur_page, the first page of index K* entries is pinned if file is not empty. Its key field is initialized to the passed key_val (a caller cannot reuse this Key object safely until the HashIndex Scanner is destroyed)

Parameters

index	IndexFile object to be scanned.
key_val	Key value on which to do the equality scan

~HashIndexScanner()

HashIndexScanner::~HashIndexScanner

(

)

Destructor.

Precondition

If the end of the Index scan has not been reached, cur_page is pinned.

Postcondition

cur_page is released.

Member Function Documentation

getNext() [1/2]

RecordId HashIndexScanner::getNext

(

)

Returns RecordId of the next matching K* in the scan.

This is the getNext method to use if you only need the RID of the next matching entry and do not also want the key value of the next matching K* entry in the scan.

Returns

Next RecordId, or INVALID_RECORD_ID if the end of the scan is reached.

getNext() [2/2]

RecordId HashIndexScanner::getNext

(

Key *

key

)

Returns RecordId of the next matching K* and sets the data part of the passed Key to the Key value of the next matching K* entry in the scan.

This version of getNext should be used when the values of the key part of the K* entry are needed by the caller (for some non-equality searches this may be the case).

Precondition

the caller passes in a Key object that has been initialized to match the Key type of the index and with appropriate data field capacity to store the largest (in num bytes) key value (this function doesn't check capacity) TODO: should this function test if passed Key object is formatted correctly to store copy of Key value??

Postcondition

RecordId of the next Record is returned and data of the passed Key object is initialized to that of the next Key part of the next K* entry in the scan. If there are no more records in the Index scan, INVALID_RECORD_ID is returned.

Returns

Next RecordId, or INVALID_RECORD_ID if the end of the scan is reached.

Member Data Documentation

bucket_scanner

HashBucketPageScanner* HashIndexScanner::bucket_scanner

private

Scanner over a hash bucket page

buf_mgr

BufferManager* HashIndexScanner::buf_mgr

private

Pointer to the BufferManager of SwatDB (for retrieving/pinning pages).

cur_dir_pid

PageId HashIndexScanner::cur_dir_pid

private

PageId of the Current Directory Page (useful for the full index scan)

cur_page

HashBucketPage* HashIndexScanner::cur_page

private

Hash Bucket object that contains the latest Record scanned. if this is equal to nullptr then the scan is complete

cur_pid

PageId HashIndexScanner::cur_pid

private

PageId of the Page that contains the latest Record scanned. if this is equal to INVALID_PAGE_ID then the scan is complete.

full_scan

bool HashIndexScanner::full_scan

private

set to true if this is a full scan of the index

index

HashIndexFile* HashIndexScanner::index

private

Pointer to the HashIndexFile to be scanned.

scan_complete

bool HashIndexScanner::scan_complete

private

set to true if the scan is done

search_key

Key* HashIndexScanner::search_key

private

The search Key value on which to scan for matches. Because hash indexes don't support range scans, and only support EQUAL Comp operators, we only need to store the search Key.

This stores a copy of the value passed to the constructor.

---

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

• hashindexscanner.h