I have my criterias separately and should be part of the cache. Also my objects have a lot of relations, so I can’t quickly set the cache limit to just 2, as it might load a bunch of relations on multiple levels.
I’m not sure what you mean by that. Do you mean that you are going to retrieve some criteria from cache?
I think that we can set the cache limit to 2 as long as the eager loading is concerned no matter what complicated relations we might have. But, yeah, I agree that the lazy loading could be a headache.
public function getEditionQuestions($edition_id) {
// Warning: Please modify the following code to remove attributes that
// should not be searched.
$criteria = new CDbCriteria;
$criteria->compare('edition_id', $edition_id, true);
return new CActiveDataProvider($this, array(
'criteria' => $criteria,
'sort' => array(
'defaultOrder' => 't.`order` ASC',
)
));
}
OK. It’s very easy. You just have to follow the wiki.
public function getEditionQuestions($edition_id) {
$criteria = new CDbCriteria;
$criteria->compare('edition_id', $edition_id, true);
$dependecy = new CDbCacheDependency('SELECT MAX(update_time) FROM tbl_question');
$duration = 60 * 60 * 24; // 1 day
return new CActiveDataProvider(Question::model()->cache($duration, $dependency, 2), array(
'criteria' => $criteria,
'sort' => array(
'defaultOrder' => 't.`order` ASC',
)
));
}
In the above I’m assuming your model is ‘Question’ and your table is ‘tbl_question’ that has ‘update_time’ column.
It’s true that the criteria for this data provider is dynamically constructed. But it’s OK.
For instance, when you call this method with $edition_id set to ‘1234’, it will create a set of 2 cache IDs from the actual sqls: 1 for the count of rows and the other for the rows(results).
And when you call it with $edition_id set to ‘5678’, it will then create ANOTHER set of IDs.
The number of cache IDs will count up by 2 with each different $edition_ids.
In each case, if CActiveDataProvider has found an entry with the specified ID in the cache, and it is not outdated by the duration or the dependency, it will return the cached result. Otherwise it will execute the query and store the result in the cache before returning it.
I will try the above code, but I am not sure how you set the cache id, and where do you tell that each different $edition_id should be on different cache.
We set the ID for a cache entry by a string. The cache component will then generate an unique key for the given string and use it to identify the cache entry.
If given the same string, then it will be the same entry.
In query caching, the sql statement is treated as the ID string.
Query caching is sometimes very expensive because of that.
If you have added a new record or modified a record, then all of them will become invalid by the specified dependency. … If it will not become invalid, it will be a disaster.
Each cache entry has its own duration and dependency. The validity of the cache entry will be evaluated with its own duration and dependency when it gets accessed.
If the cache entry for page #1 has become invalid because of the duration, then that of page #2 will become invalid sooner or later, because they have the same duration and probably has been created almost at the same time.
If the cache entry for page #1 has become invalid because of the dependency change, then that of page #2 will become invalid exactly at the same time.
Please read the guide for the meaning of ID, duration and dependency of the cache.
Just a quick question - would you happen to know how to invalidate (flush) the cache if a record is deleted? I imagine I could add a CDbCacheDependency to monitor the row count on the database table (i.e.: the row count decreases by one upon deleting a record), but I’d rather be able to programmatically do the invalidation. Any thoughts?
