Re: svn: /phd/trunk/phpdotnet/phd/Reader/ Partial.php
| From: | Hannes Magnusson | Date: | Sun, 24 Apr 2011 10:34:17 +0000 |
| Subject: | Re: svn: /phd/trunk/phpdotnet/phd/Reader/ Partial.php | ||
| References: | 1 2 | Groups: | php.doc.cvs |
| Request: | Send a blank email to doc-cvs+get-7999@lists.php.net to get a copy of this message | ||
On Sun, Apr 24, 2011 at 10:58, Richard Quadling <rquadling@gmail.com> wrote:
> On 24 April 2011 09:25, Hannes Magnusson <bjori@php.net> wrote:
>> bjori Sun, 24 Apr 2011 08:25:29 +0000
>>
>> Revision:
>> http://svn.php.net/viewvc?view=revision&revision=310450
>>
>> Log:
>> shave off few seconds when seeking for an chunk to render
>>
>> Changed paths:
>> U phd/trunk/phpdotnet/phd/Reader/Partial.php
>>
>> Modified: phd/trunk/phpdotnet/phd/Reader/Partial.php
>> ===================================================================
>> --- phd/trunk/phpdotnet/phd/Reader/Partial.php 2011-04-24 07:11:34 UTC (rev 310449)
>> +++ phd/trunk/phpdotnet/phd/Reader/Partial.php 2011-04-24 08:25:29 UTC (rev 310450)
>> @@ -6,6 +6,7 @@
>> {
>> protected $partial = array();
>> protected $skip = array();
>> + protected $parents = array();
>>
>> public function __construct() {
>> parent::__construct();
>> @@ -28,6 +29,22 @@
>> } else {
>> throw new \Exception("Didn't get any IDs to seek");
>> }
>> + $sqlite = new \SQLite3(Config::output_dir() . "index.sqlite");
>> +
>> + // Fetch all ancestors of the ids we should render
>> + foreach($render_ids as $p => $v) {
>> + do {
>> + $id = $sqlite->escapeString($p);
>> + $row = $sqlite->query("SELECT parent_id FROM ids WHERE
>> docbook_id = '$id'")->fetchArray(SQLITE3_ASSOC);
>> + if ($row["parent_id"]) {
>> + $parents[] = $p = $row["parent_id"];
>> + continue;
>> + }
>> + break;
>> + } while(1);
>> + }
>> +
>> + $this->parents = $parents;
>> }
>>
>> public function read() {
>> @@ -92,6 +109,13 @@
>> } elseif (empty($this->partial)) {
>> return false;
>> } else {
>> + if ($id) {
>> + // If this id isn't one of our ancestors we can jump
>> + // completely over it
>> + if (!in_array($id, $this->parents)) {
>> + parent::next();
>> + }
>> + }
>> $ignore = true;
>> }
>> }
>
> The approach being used here (the Adjacent Set Model) requires a query
> to be executed for each level until you run out of levels.
>
> For very large numbers of levels, there is a much more efficient
> mechanism, the Nested Set Model.
>
>
> http://dev.mysql.com/tech-resources/articles/hierarchical-data.html
>
> For any number of levels, a single query will return all the data.
>
> Is this something worth considering?
No. Those queries (even in the worst case, 6*partial-id-count queries)
take a fraction of a nanosecond.
Besides, if we want to go for a proper hierarcical data model on the
database level we need to store a lot more information, which will
increase the save and load time by more then it takes to run those 6
queries.
Keeping in mind these queries are only executed during partial build..
there is no way its worth adding more data to the db :)
-Hannes