db5.c

Demonstrates the use of duplicate items; reads a text file from stdin and creates a Database entry for every word, even if the word is a duplicate. Then prints all words in alphabetical order and prints the line number of each occurrence.

#include <stdio.h>
#include <string.h>
#include <ham/hamsterdb.h>

int 
main(int argc, char **argv)
{
    ham_status_t st;      /* status variable */
    ham_db_t *db;         /* hamsterdb database object */
    ham_cursor_t *cursor; /* a database cursor */
    char line[1024*4];    /* a buffer for reading lines */
    unsigned lineno=0;    /* the current line number */
    ham_key_t key;
    ham_record_t record;

    memset(&key, 0, sizeof(key));
    memset(&record, 0, sizeof(record));

    printf("This sample uses hamsterdb and duplicate keys to list all words "
            "in the\noriginal order, together with their line number.\n");
    printf("Reading from stdin...\n");

    /*
     * first step: create a new hamsterdb object 
     */
    st=ham_new(&db);
    if (st!=HAM_SUCCESS) {
        printf("ham_new() failed with error %d\n", st);
        return (-1);
    }

    /*
     * second step: create a new database with support for duplicate keys
     *
     * we could create an in-memory-database to speed up the sorting.
     */
    st=ham_create(db, "test.db", HAM_ENABLE_DUPLICATES, 0664);
    if (st!=HAM_SUCCESS) {
        printf("ham_create() failed with error %d\n", st);
        return (-1);
    }

    /*
     * now we read each line from stdin and split it in words; then each 
     * word is inserted into the database
     */
    while (fgets(line, sizeof(line), stdin)) {
        char *start=line, *p;
        lineno++;

        /*
         * strtok is not the best function because it's not threadsafe
         * and not flexible, but it's good enough for this example.
         */
        while ((p=strtok(start, " \t\r\n"))) {
            key.data=p;
            key.size=(ham_size_t)strlen(p)+1; /* also store the terminating 
                                                 0-byte */
            record.data=&lineno;
            record.size=sizeof(lineno);


            /* note: the second parameter of ham_insert() is reserved; set it 
             * to NULL; the flag HAM_DUPLICATE inserts a duplicate key */
            st=ham_insert(db, 0, &key, &record, HAM_DUPLICATE);
            if (st!=HAM_SUCCESS) {
                printf("ham_insert() failed with error %d\n", st);
                return (-1);
            }
            printf(".");

            start=0;
        }
    }

    /* 
     * create a cursor 
     */
    st=ham_cursor_create(db, 0, 0, &cursor);
    if (st!=HAM_SUCCESS) {
        printf("ham_cursor_create() failed with error %d\n", st);
        return (-1);
    }

    /*
     * iterate over all items and print them
     */
    while (1) {
        st=ham_cursor_move(cursor, &key, &record, HAM_CURSOR_NEXT);
        if (st!=HAM_SUCCESS) {
            /* reached end of the database? */
            if (st==HAM_KEY_NOT_FOUND)
                break;
            else {
                printf("ham_cursor_next() failed with error %d\n", st);
                return (-1);
            }
        }

        /* 
         * print the word and the line number
         */
        printf("%s: appeared in line %u\n", (const char *)key.data, 
                *(unsigned *)record.data);
    }

    /*
     * then close the database handle; the flag
     * HAM_AUTO_CLEANUP will automatically close all cursors, and we
     * do not need to call ham_cursor_close
     */
    st=ham_close(db, HAM_AUTO_CLEANUP);
    if (st!=HAM_SUCCESS) {
        printf("ham_close() failed with error %d\n", st);
        return (-1);
    }

    /*
     * delete the database object to avoid memory leaks
     */
    ham_delete(db);

    /* 
     * success!
     */
    return (0);
}