Randy J Ray


WebService::ISBNDB::Iterator - Iterator class for large result-sets


    # The search() method of API-derived classes returns an Iterator
    $iter = WebService::ISBNDB::API->search(Books =>
                                            { author =>
                                             'poe_edgar_allan' });

    print $iter->get_total_results, " books found.\n";
    while ($book = $iter->next)
        print $book->get_title, "\n";

    # Reset the iterator

    # Do something else with all the elements found by the search
    for ($iter->all)


This class provides an iterator object to abstract the results from a search. Searches may return anywhere from no matches to thousands. Besides the fact that trying to allocate all of that data at once could overwhelm system memory, the isbndb.com service returns data in "pages", rather than risk sending an overwhelming response.

The iterator stores information about the initial request, and as the user progresses past the in-memory slice of data, it makes subsequent requests behind the scenes to refresh the data until the end of the results-set is reached.

It is not expected that users will manually create iterators. Iterators will be created as needed by the search method in the API classes.


Methods are broken in the following groups:



The constructor is based on the Class::Std model. The argument it takes is a hash-reference whose key/value pairs are attribute names and values. The attributes are defined below, in "Accessor Methods".

The only required attributes in the arguments list are request_args and contents. The first is the set of arguments used in the initial request made to the service. They are reused when subsequent pages need to be fetched. The second is the initial set of objects, fetched from the first page of results.

Iterator Methods

These methods are the general-use interface between the user and the iterator. In most cases, an application will only need to use the methods listed here:


Return the first element in the results-set. Regardless of the current position within the iterator, this is always the very first element (or undef, if there were no elements found by the search). This does not alter the position of the internal pointer, or trigger any additional requests to the data source.


Return the next element off the iterator, or undef if the iterator is exhausted. All elements returned by an iterator descend from WebService::ISBNDB::API. All elements in a given iterator will always be from the same implementation class. The iterator does not explicitly identify the class of the objects, since the application had to have had some degree of knowledge before making the call to search.


Returns the full set of results from the iterator, from the beginning to the end (if the iterator has already been read some number of times, it is reset before the list is constructed). The return value is the list of elements when called in a list-context, or a list-reference of the elements when called in a scalar context. The iterator will be in an exhausted state after this returns.


Resets the internal counter within the iterator to the beginning of the list. This allows the iterator to be re-used when and if the user desires.


When a request (via next()) goes past the internal set of data, this method is called to request the next page of results from the data source, until the last page has been read. This method alters the page_number, contents and shown_results attributes. If the user has set a hook (via set_fetch_page_hook()), it is called with the arguments for the request just prior to the request itself. The arguments are those provided in the request_args attribute, plus a page_number argument set to the page that is being requested.

Accessor Methods

The accessor methods provide access to the internal attributes of the object. These attributes are:


The total number of results in the result-set, not to be confused with the number of results currently in memory.


The size of the "page" returned by the data source, in turn the maximum number of elements held internally by the iterator at any given time. As the index proceeds to the end of the in-memory list, a new page is fetched and this many new elements replace the previous set internally.


The number of the page of results currently held within the iterator. When the iterator fetches a new page, this is incremented. When the iterator is reset, this is set to 1.


The number of results currently held within the iterator. When the last page of a results-set is fetched, it may have fewer than page_size elements in it. This attribute will always identify the number of elements currently kept internally.


The list reference used internally to store the current set of objects for the page of results held by the iterator. Be careful with this value, as changing its contents can change the internal state of the iterator.


The hash reference that stores the original request arguments used to fetch the initial page of data from the data source. This is used to make any additional requests for subsequent pages, as needed. Be careful with the value, as changing its contents can affect the iterator's ability to fetch further pages.


The integer value that marks the current position within the iterator. The value is the position within the whole set of results, not just within the single page held internally.


The WebService::IDBNDB::Agent instance that is used to fetch additional pages as needed. It is generally set at object-construction time by the API object that creates the iterator. If it is not specified in the constructor, the get_default_agent method of WebService::ISBNDB::API is called.


If this attribute has a value that is a code-reference, the code-reference is invoked with the arguments that are going to be passed to the request method of the agent. The hook (or callback) will receive the iterator object referent and the hash-reference of arguments, as if it had been called as a method in this class. The arguments are those stored in request_args as well as one additional argument, page_number, containing the number of the page being requested.

Note that the hook will not be called for the first page fetched from the data source. That is because that fetch is done outside the scope of the iterator class, and the data from that initial fetch is provided when the iterator is constructed.

Note that for most of the attributes, only the "get" accessor is documented. Users should not need to manually set any of the attributes (except for fetch_page_hook) unless they are sub-classing this class:


Return the relevant attribute's value. Note, again, that get_contents() and get_request_args() return the actual reference value used internally. Changes to the contents of those reference values may impact the behavior of the iterator itself.


Set a hook (callback) to be called each time the iterator has to fetch a new page from the data source. The value is a code-reference, and is called with the iterator object and a hash-reference of the request arguments as parameters. Any return value is ignored. If the hook dies, an exception is thrown by fetch_next_page() with the error message.


Get the current hook value, if any.


WebService::ISBNDB::API, Class::Std


Randy J. Ray <rjray@blackperl.com>


This module and the code within are released under the terms of the Artistic License 2.0 (http://www.opensource.org/licenses/artistic-license-2.0.php). This code may be redistributed under either the Artistic License or the GNU Lesser General Public License (LGPL) version 2.1 (http://www.opensource.org/licenses/lgpl-license.php).