Indexing

Indexing is a way to organize records in a file

• index file - is file containing indexes containing a key and a pointer to the corresponding record in the data file

 

struct Index{

KEY key;

long int record;

};

• data file - contains data records

 

 

Files in C

C supports two types of files:

• character files (streams)
• binary files

Operations on files:

• open a file:

FILE* fopen (char * name, char mode)

• close a file

void fclose (FILE* file)

• read a file

size_t fread (void* b, size_t s, size_t n, FILE* f)

• write a file

size_t fwrite (void* b, size_t s, size_t n, FILE* f)

 

• position a file

int fseek (FILE* f, long offset, int base)

 

base can be:

SEEK_SET - from the beginning

SEEK_CUR - from current position

SEEK_END - from the end

• test end of file

int feof (FILE* f)

 

Example Create a file containing records of the following form:

struct Records{

char first[20] ;

char last[20];

int id_num; // this is the key

int age;

}

together with an index file

Linear indexing

• the index file is organized as a sequence of indexes sorted by the keys

• the disadvantages are the same as for array-implementation of lists
• it is possible to extend this indexing scheme to multiple key

 

Tree Indexing

• use BST's to organize indexes to obtain efficient search, insertion and deletion

• problem: BST may become unbalanced ... for files this is a greater problems than for data structures stored in memory

2-3 Trees

Definition. 2-3 tree:

q a node contains one or two keys

q every internal node contains two or three children

q the leaves are all at the same level in the tree

q the tree has the search property:

• nodes in the left subtree < value of the first key
• nodes in the middle subtree >= value of the first key
• node in the right subtree >= value of the second key (if there is any)

structure TTNode{

int nkeys;

KEY lkey;

KEY rkey;

TTNode* left;

TTNode* center;

TTNode* right;

};

 

Searching in a 2-3 tree

• similar to binary search, except for the three-way branching

Inserting and deleting in a 2-3 tree

• do not create new nodes at the bottom of the tree - they use split or merge and have effect higher in the tree

 

Insertion

B-Trees

Definition. A B-Tree of order m is a tree with the following properties:

1. The root is either a leaf, or it has at least two children
2. Each node, except for the root has between m/2 and m children.
3. All leaves are at the same level in the tree, so that the tree is always height balanced.

B+-Trees

• a B-Tree in which only the leaf nodes contain data (all the other nodes only contain keys)