WordPress Custom Select Query - mysql

I really don't know enough about MySQL queries and it's showing.
I have a custom field set for every post. The custom field stores the posts source URL in a key called "source_url".
I have it working with the below WP_Query parameters, but it's incredibly slow. Keep in mind it's possible to 50+ urls to search for.
So, given an array of source URL's, I want to fetch the matching posts.
For example, here is what I currently have that's slow in WP_Query:
// var_dump of $urls array (this could be 50+ urls)
array(7) {
[0]=>
string(42) "http://www.youtube.com/watch?v=FMghvnqDhT8"
[1]=>
string(42) "http://www.youtube.com/watch?v=RY-yUFpXTnM"
[2]=>
string(58) "http://www.youtube.com/watch?v=nIm2dnyJ1Ps&feature=related"
[3]=>
string(42) "http://www.youtube.com/watch?v=NoCtRQlJAqM"
[4]=>
string(57) "http://holidaycustoms.blogspot.com/2012/08/busy-week.html"
[5]=>
string(42) "http://www.youtube.com/watch?v=DcZvg197Ie4"
[6]=>
string(42) "http://www.youtube.com/watch?v=7P3UEbLmLuo"
}
// Build Media Query
$meta_query = array(
'relation' => 'OR'
);
foreach( $urls as $url ) {
$meta_query[] = array(
'key' => 'source_url',
'value' => $url
);
}
// Get 20 matching posts from a category set by a variable
$args = array(
'post_type' => 'post',
'posts_per_page' => 20,
'orderby' => 'rand',
'cat' => $cat_ID,
'meta_query' => $meta_query
);
$posts = get_posts($args);
What I'm looking to do is replace the above code with a custom query select, which I have read is much faster than WP_Query.
But I don't know enough about MySQL or the WP database to build the custom select query. Can anyone help? Thanks in advance!

In the post you linked yourself, the first reply already states that
[...] the default schema doesn't even have an index on the value column
Which is far more severe a problem than any you would have with a query generator, because without an index the DBMS has to traverse the whole table and compare strings of each field.
Adding an index is fairly easy with an appropriate management tool like PHPMyAdmin. The offending table you will need to add an index to is called wp_postmeta and the field that needs an index is meta_value, and the index type should be INDEX.
Adding an index is transparent and does not affect wordpress other than in performance. It could take some time though since, well MySQL needs to traverse the whole table. Also, because you are indexing string data, the index will be quite big.
You should also try using appropriate structures for your query. You are currently using a big ORed selection with different values but always the same field. There is a construct for just that, and it's called IN.
...
// Build Media Query
$meta_query = array();
$meta_query[] = array(
'key' => 'source_url',
'value' => $urls,
'compare' => 'IN'
);
// Get 20 matching posts from a category set by a variable
..
(Untested. I actually never did this, Reference)
The performance gain would be negligible compared to adding an index I assume, but your code would become a lot simpler.

Related

Checking 2 Column for Duplicate

Currently, I have a system to hold main data
1) The email
2) The owner(user_id)
Every time someone uploads , I need to make sure that it doesn't not exist in the system. The catch is as I upload more and more, the amount of time taken to check for duplicate will grow steeply, just like the graph as shown.
Question
1) How do i check for duplicate efficiently?
2) I indexed the user_id and the email should I Fulltext it? I wont be reading the text but will be searching for it as a whole, so index is more logical?
3) I also read about creating Hash combining email&owner id then index the hash. Will it be a big difference from the current method?
4) Last method i thought of was to create a primary key for both email and user_id , once again idk how the performance would turn out.
Please advice.
Code
$exist = DB::table('contact')->where('email', $row['email'])->where('user_id', $user_id)->count();
if($exist < 1){
DB::table('contact')->insert(
['email' => $row['email'], 'name' => $row['name'], 'user_id' => $user_id]
);
}
Use Laravel Validator:
public function store(Request $request)
{
$this->validate($request, [
'user_id' => 'required|unique',
'email' => 'required|unique',
]);
//some logic here
}
Also you should use unique constraint in your database.

CakePHP 2.3: Conditionally retrieve unique model information based on another model's use of it

I'm trying to create an AJAX form whereby the content of a select field populates based on the choice of a preceding select field (you see this a lot with 'country' populating 'state/province'). In my case, I want users to be able to choose their province only if active accounts exist in it.
The Javascript I can write no problem. Fetching the data is where I'm... not so much stuck as doing too much work. CakePHP likes to build select fields with options in an array of the form
$options = array(select_option_value => display_text)
My strategy, though functional, must be more convoluted than cake intended (this a is segment of a controller method).
$provinceData = $this->Account->find('all',array('recursive' => 0,
'joins' => array(
array(
'table' => 'provinces',
'type' => 'LEFT',
'conditions' => array('Account.province_id = provinces.id')
)),
'fields'=>array('provinces.id', 'provinces.name', 'provinces.abbrev'),
'conditions' => array('registration > 2')));
$provinces = array();
foreach($provinceData as $pd) {
/*note: lowercase, plural below b/c can't get 'alias' => 'Province'
to work in joins array above : ( */
$id = $pd['provinces']['id'];
$name = $pd['provinces']['name'];
$provinces[$id] = $name;
}
$this->set(compact('provinces'));
Can anyone point out a more appropriate way to do this? I assume there must be a MySQL query that can do this, but I'm pretty bad at writing elaborate MySQL queries in the first place, let alone via Cake's convention (and, for you MySQL gurus out there, I'm happy to do this from a Model->query(//MySQL code) call instead!
Any and all help truly appreciated.
Assuming the relationship Account belongsTo Province you can try this code:
$accounts = $this->Account->find(
'all',
array(
'fields' => array('Account.province_id', 'Province.name'),
'conditions' => array('Account.registration > 2'),
'group' => 'Account.province_id'
)
);
$provinces = Hash::combine($accounts, '{n}.Account.province_id', '{n}.Province.name');
$this->set(compact('provinces'));
edit: missed bracket and a period instead of an underscore . Now should work

CakePHP2.3: Building complex conditional Model->find() options

I'm not savvy with MySQL or databases generally, so here's a model of my data (table{cols}]) in order to make my question coherent:
Domains{id, name} Note: 'domains' here does not refer to web domains
Subdomains{id, domain_id, name}
Items{id, subdomain_id, name}
SubdomainsItems{id, subdomain_id, item_id} no domain_id column!
My Items Controller has a function, fetchWithin($domains, $subdomains) which, ultimately, should just execute one of two complexish find(). It's the complexish I can't get past.
Programmatically I can achieve this, but I'm quite certain the better way is by clever joins and the like. Alas, currently this is approach:
If $domainsis empty, do only steps 2&3, otherwise:
foreach($domains as $d): get all the rows of Subdomains where Subdomain.domain_id = Domains.id as $subdomains
foreach($subdomains as $s) : go get all the rows of SubdomainsItems where SubdomainsItems.subdomain_id = Subdomains.id as $item_ids
foreach($items_ids as $i): get all the rows of Items where Items.id = SubdomainsItems.items_id
This works, but I think this is obviating the power of a relational database and I'd like to understand how this should be done (ie. according to either Cakephp convention or simply by whatever MySQL statement would achieve this).
Help would be hugely appreciated, I try to learn the more complex aspects of SQL but it just goes right over my head. :S
Understanding the necessary query
With the structure described in the question the kind of query necessary is of the form:
SELECT
*
FROM
items
LEFT JOIN
subdomains ON (
items.subdomain_id = subdomains.id
)
LEFT JOIN
domains ON (
subdomains.domain_id = domains.id
)
WHERE
domains.name = "foo"
AND
subdomains.name IN ('some', 'list', 'of', 'subdomains');
Compared to the logic in the question this joins all three tables together and permits finding all items by domain name, or subdomain name (or any other criteria involving any or all three tables); Generally speaking if you want to find data in a db and use more than one query to get it - there's a more efficient way to do it.
Implementing the find call
There are a number of ways of creating such a query with Cake. The simplest, probably, is to use the join key and just specify the joins explicitly:
function fetchWithin($domains = null, $subdomains = null) {
$params = array(
'joins' => array(
array('table' => 'subdomains',
'alias' => 'Subdomain',
'type' => 'LEFT',
'conditions' => array(
'Subdomain.id = Item.subdomain_id',
)
),
array('table' => 'domains',
'alias' => 'Domain',
'type' => 'LEFT',
'conditions' => array(
'Domain.id = Subdomain.domain_id',
)
)
)
);
if ($domains) { // single value or an array
$params['conditions']['Domain.name'] = $domains;
}
if ($subdomains) { // single value or an array
$params['conditions']['Subdomain.name'] = $subdomains;
}
return $this->find('all', $params);
}

How to get return only specified fields with a find() / query

Currently I am using
$posts = $this->Post->find('all');
and it returns complete entries of the table posts but how can I get only partial entries. Some thing like
SELECT body,title FROM posts
Pass the "fields" option and specify (in an array) what fields you want to retrieve.
$posts = $this->Post->find('all', array(
'fields' => array('body', 'title')
));
I hate to be the guy that writes this, cause it usually bothers me, but I think this question is worth of "RTM".
This page in the book explains it all: http://book.cakephp.org/2.0/en/models/retrieving-your-data.html

CakePHP: finding information in distantly related models

I have a News model, which has a HABTM relationship with an Artists model, and an artist in turn hasMany tourdates.
If I want to find all tourdates related to the current news item, what is an efficient way of phrasing that for CakePHP?
This is what I have so far; I'm wondering if (a) it looks like it should work, and (b) if there's any more concise way of writing it:
$relatedartists = $this->News->ArtistsNews->find('list', array(
'conditions'=>array('ArtistsNews.news_id' => $id),
'fields'=>array('artist_id')
));
$livedates = $this->News->Artists->Tour->find('all', array(
'conditions'=>array('Tour.artist_id'=> $relatedartists,
'date >= ' . time()),
'order'=>'date ASC'
));
What you have is pretty good. I always prefer to use multiple queries rather than use massive joins which create temporary tables. It can reduce performance somewhat.
You might also try something like the below
$opts = array(
'conditions' => array(
'ArtistsNews.news_id' => $id
)
);
$this->News->Artists->recursive = 2;
$this->News->Artists->find('all', $opts);
Something along the likes of this query will also get you what you need (haven't error checked)