I am trying to fix an issue with our website and I'm afraid it is a bit out of my depth at the moment. We have a social network website and users can search for friends through a friend section.
The problem is that when you open that section the SELECT query seems to run a search for ALL users on the site. And we have over 15,000 users so the page is incredibly slow and normally times out.
I was able to limit the displayed result to 15 users per page, but the query is not well optimized and apparently selecting all users.
Here is the SELECT query code we have at the moment:
$sql_sel_friend_req = "SELECT m.*, mp.iMAgeGroup as modelAgeGroup,
c.iMAgeGroup as clientAgeGroup, mo.iModelId, mp.vImage as modelImage,
cc.vImage as clientImage FROM member m
left join model as mo on (mo.iModelId = m.MemberId && m.MemberType = 'Model')
left join model_profile as mp on (mp.iModelId = mo.iModelId)
left join client as c on (c.iClientId = m.MemberId && m.MemberType = 'Client')
left join client_company as cc on (cc.iClientId = c.iClientId) WHERE m.MemberId != " . $_SESSION['sess_Login_Id'] . " && m.MemberStatus = 'Active' &&
m.MemberId not in (SELECT distinct(mf.iFriendMemberId1) from member_friends mf where mf.iFriendMemberId2 = " . $_SESSION['sess_Login_Id'] . " && mf.eFriendMemberType2 = '" . $_SESSION['sess_Login_Type'] . "') &&
m.MemberId not in (SELECT distinct(mf1.iFriendMemberId2) from member_friends mf1 where mf1.iFriendMemberId1 = " . $_SESSION['sess_Login_Id'] . " && mf1.eFriendMemberType1 = '" . $_SESSION['sess_Login_Type'] . "') ";
I've tried to use Limit and Offset but this only caused SQL errors.
If you have any ideas please let me know.
Thank you
Related
I'm trying to get the results from 'comments' table with the exception of the ones in the 'readsbaby' table.
The result i'm getting is from comments but no effect of the NOT EXISTS statement, so the result is all the comments.
Both tables have common data that should not be included in the result.
I checked the data and the syntax many times.
Still this query will return all comments without taking in consideration the AND NOT EXISTS close.
public function get_user_comments($post_id)
{
$user_id = $this->session->userdata('id');
$group_id = $this->session->userdata('group_id');
$sql = "SELECT *
FROM comments
WHERE DATE(created_on) > DATE_SUB(CURDATE(), INTERVAL 1 DAY)
AND comments.group_id = " . $group_id . "
AND comments.user_id != " . $user_id . "
AND NOT EXISTS ( SELECT *
FROM readsbaby
WHERE comments.id = readsbaby.notification_id
AND comments.group_id = readsbaby.group_id
AND readsbaby.user_id = " . $this->session->userdata('id') . "
AND comments.nature1 = readsbaby.notification_type
) ";
return $data=$this->db->query($sql)->result_array();
}
Expecting to get the result filtered by the NOT EXISTS close.
I'd change the query this way, to have only the rows without any match in the table readsbaby:
public function get_user_comments($post_id)
{
$user_id = $this->session->userdata('id');
$group_id = $this->session->userdata('group_id');
$sql = "SELECT *
FROM comments
LEFT OUTER JOIN readsbaby ON comments.id = readsbaby.notification_id
AND comments.group_id = readsbaby.group_id
AND readsbaby.user_id = " . $this->session->userdata('id') . "
AND comments.nature1 = readsbaby.notification_type
WHERE DATE(created_on) > DATE_SUB(CURDATE(), INTERVAL 1 DAY)
AND comments.group_id = " . $group_id . "
AND comments.user_id != " . $user_id . "
AND ISNULL(readsbaby.notification_id )";
return $data=$this->db->query($sql)->result_array();
}
This is easier to accomplish using the correct mysql JOIN.
From your post, I have gathered that you want to retrieve all the values from comments where they don't exist in readsbaby. The id column in comments associates with the notification_id in readsbaby for any 'matching' (duplicate) entries.
On this assumption, you can do this at a simplistic level like so
SELECT c.* FROM comments c
LEFT JOIN readsbaby r ON c.id = r.id WHERE r.notification_id IS NULL;
...which should translate to your context as something like:
SELECT c.* FROM comments c
LEFT JOIN readsbaby r ON c.id = r.notification_id
WHERE
r.notification_id IS NULL
AND DATE(c.created_on) > DATE_SUB(CURDATE(), INTERVAL 1 DAY)
AND c.group_id = " . $group_id . "
AND c.user_id != " . $user_id . "
If, I have this wrong and you're looking to only get the data that isn't present in both, you'd need to use something like:
SELECT c.* FROM comments c
LEFT JOIN readsbaby r ON c.id = r.notification_id
WHERE
r.notification_id IS NULL
AND DATE(c.created_on) > DATE_SUB(CURDATE(), INTERVAL 1 DAY)
AND c.group_id = " . $group_id . "
AND c.user_id != " . $user_id . "
UNION
SELECT r.* FROM readsbaby r
LEFT JOIN comments c ON c.id = r.notification_id
WHERE
c.id IS NULL
AND r.group_id = " . $group_id . "
AND r.user_id != " . $user_id . "
This may need tweaking to suit, as there's very little to explain the data and table relationship.
Note: Do not inject SQL as you've done. Consider parameterising your query with PDO or any other prepared statements (e.g. http://php.net/manual/en/pdo.prepare.php)
i am trying to write a query which will help me recover class, batch, and institution details from db
so far i have managed to make the following work:
SELECT
gallery.path_url AS insimage,
r3.*
FROM
gallery
INNER JOIN (
SELECT
institution.id AS insid,
institution.ProfileImageID,
institution.`name` AS insname,
institution.Locality AS inslocality,
institution.RegistrationFee AS ins_reg_fee,
institution.IsTrail AS ins_is_trail,
institution.LatLong AS latlong,
r2., activity.`name` AS activityname
FROM
institution
INNER JOIN (
SELECT
class., count() AS batch_count,
r1.
FROM
class
INNER JOIN (
SELECT
batch.ID AS batch_id,
batch.ClassID AS class_id,
batch.LevelID AS level_id,
batch.agegroupID AS agegroup_id,
count(*) AS num_batches,
min(Price) AS min_price,
max(Price) AS max_price
FROM
batch,
batchstarttimes
WHERE
batch.ID = batchstarttimes.BatchID
AND batchstarttimes.StartTime BETWEEN '" . $sttime . "'
AND '" . $endtime . "'
GROUP BY
batch.ID
) AS r1
WHERE
r1.class_id = class.ID
GROUP BY
r1.class_id
) AS r2 ON r2.institutionid = institution.id
INNER JOIN activity ON activity.id = r2.activityid
WHERE
activity.`name` LIKE '%" . $searchterm . "%'
AND institution.Locality LIKE '%" . $locality . "%'
AND institution.StatusID = 1
AND level_id = 1
AND agegroup_id = 5
) AS r3 ON r3.ProfileImageID = gallery.ID
I am stuck when i try to add two more filters batch.LevelID and batch.agegroupID
any way i can get help or someone point me to a sql visualizr tool will be helpful.
In your WHERE clause, replace:
AND level_id = 1
AND agegroup_id = 5
with:
AND batch.LevelID = CASE WHEN " . $levelID . " = 0 THEN batch.LevelID ELSE " . $levelID . " END
AND batch.agegroupID = CASE WHEN " . $ageGroupID . " = 0 THEN batch.agegroupID ELSE " . $ageGroupID . " END
The point is to compare batch.LevelID and batch.agegroupID against themselves if you want to ignore them for the filtering (this way they'll always be true) or compare them against $levelID and $ageGroupID values for specific rows.
Also, it's not recommended to use the variables directly in the query, as it's vulnerable to SQL injections attacks. Read about prepared statements here.
I have a query and i want random result on advertisement table. Advertise table have own id but when i use RAND(advertise.id) it won't work and i don't know how it will. I am using laravel framework so if possible i can use PHP also to show random advertise result. Here is code can anyone tell me where i use RAND() Mysql function.
SELECT providers.id,
adv__managements.id AS 'adv_id',
adv__managements.images,
adv__managements.categories,
adv__managements.subcategories,
adv__managements.title,
adv__managements.description,
adv__managements.role,
adv__managements.plan_id,
userinformation.user_id,
userinformation.zip_code,
plans.active
FROM userinformation INNER JOIN providers ON userinformation.user_id = providers.user_id
INNER JOIN adv__managements ON adv__managements.range = providers.range
INNER JOIN plans ON plans.user_id = userinformation.user_id
where GetDistance('km'," . doubleval($info->lat) . ", " . doubleval($info->lon) .", " . doubleval(\Auth::user()->userinfo->latitude) . ", " . doubleval(\Auth::user()->userinfo->longitude) . ") < providers.range AND plans.active = 1 LIMIT 3
And can anyone tell how to convert this query into laravel ?
So I have 2 tables:
user_selection (id, user_id, fiche_id)
fiche (id, title, ...)
(fiche = sheet)
I have a research request which gives me some fiche ids
SELECT fiche.*, IF(fiche.fiche_type_id = 2,fiche.instance,fiche_type_texte.type_texte) AS type_texte, IF(fiche.fiche_type_id = 2,CONCAT(fiche.affaire," - ", fiche.titre),fiche.titre) AS titre, IF(fiche.fiche_type_id = 1,fiche.date_publication,fiche.date_texte) AS date_publication
FROM enviroveille_bascule.fiche
LEFT JOIN enviroveille_bascule.fiche_type_texte ON(fiche.fiche_type_texte_id = fiche_type_texte.id)
LEFT JOIN enviroveille_bascule.fiche_x_keyword ON(fiche_x_keyword.fiche_id = fiche.id)
LEFT JOIN enviroveille_bascule.fiche_x_theme ON(fiche_x_theme.fiche_id = fiche.id)
LEFT JOIN enviroveille_bascule.fiche_echeance ON(fiche_echeance.fiche_id = fiche.id)
LEFT JOIN enviroveille_bascule.fiche_x_activite ON(fiche_x_activite.fiche_id = fiche.id) LEFT JOIN enviroveille_bascule.fiche_x_nomenclature ON(fiche_x_nomenclature.fiche_id = fiche.id)
WHERE 1 = 1
AND fiche_echeance.fiche_echeance_type_id IN(2)
GROUP BY fiche.id
I'd like to know if all results from my research query is in the user selection.
So I tried :
// fiche_ids is an array of fiche.id resulting from research request
SELECT COUNT(*) AS nb
FROM " . $this->bdd . $this->table_user_selection . "
WHERE user_id = '" . $user->datas_user['id'] . "'
AND fiche_id IN ('" . implode("','", $fiche_ids) . "')
if($ds['nb'] == count($fiche_ids) && count($fiche_ids) > 0) {
return true;
} else {
return false;
}
That works well
Problem is that I have some research request giving 10K+ results.
And I have to do it several time on the same page, which makes server lags.
Is there a easy way to know if all my results are in my selection?
Note that selection may contain more fiche_id than the research result.
The best would be to be able do this in one SQL request.
I have a small database, holding the details of just under 400 ponies. I wish to query that table and return a table showing the pertinant details of each pony, and it's owner's and breeder's names. The data is held primarily like so:
profiles - a table holding all info assigned to each individual pony, including it's sire's and dam's reg numbers, and it's owner's and breeder's DB assigned id's.
contacts - a table for the people's info. Joined as 'owner' and again as 'breeder' in the query below.
prm_* - multiple parameter tables, holding broad details such as colour, breed, etc.
Where I am running into trouble is when trying my first self join: querying the profiles table three times in order to retrieve the names of the sire and dam for each profile, as well as the pony's own name to begin with. When I run the query, it returns duplicate rows for many (not all) profiles. Using DISTINCT eliminated most of these, but the issue remains with the non-identical results, particularly for those ponies where no sire or dam is on record.
I have googled the problem, and it does appear here and there, but I cant quite grasp what happening in the solutions given. I'm not even certain why the problem occurs at all. Can someone please step me through the issue and the solving of it? I'd be most grateful.
My query as it stands (returns 408 results, from only 387 ponies!):
include 'conn.php';
?>
<table class="admin-display">
<thead><tr><th>No:</th><th>Name:</th><th>Sire:</th><th>Dam:</th><th>Age:</th><th>Colour:</th><th>Gender:</th><th>Owner:</th><th>Breeder:</th></tr></thead>
<?php
$i=1;
$sql = mysql_query("SELECT DISTINCT p.ProfileID, p.ProfileOwnerID, p.ProfileBreederID, p.ProfilePrefix, p.ProfileSireReg, p.ProfileDamReg,
p.ProfileGenderID, p.ProfileAdultColourID, p.ProfileColourModifierID, p.ProfileYearOfBirth,
p.ProfileYearOfDeath, p.ProfileLocalRegNumber, p.ProfileName,
sire.ProfileName AS sireName, sire.ProfilePrefix AS sirePrefix,
dam.ProfileName AS damName, dam.ProfilePrefix AS damPrefix,
owner.ContactFirstName AS owner_fname, owner.ContactLastName AS owner_lname,
breeder.ContactFirstName AS breeder_fname, breeder.ContactLastName AS breeder_lname,
BreedGender, BreedColour, BreedColourModifier
FROM profiles AS p
LEFT JOIN profiles AS sire
ON p.ProfileSireReg = sire.ProfileLocalRegNumber
LEFT JOIN profiles AS dam
ON p.ProfileDamReg = dam.ProfileLocalRegNumber
LEFT JOIN contacts AS owner
ON p.ProfileOwnerID = owner.ContactID
LEFT JOIN contacts AS breeder
ON p.ProfileBreederID = breeder.ContactID
LEFT JOIN prm_breedgender
ON p.ProfileGenderID = prm_breedgender.BreedGenderID
LEFT JOIN prm_breedcolour
ON p.ProfileAdultColourID = prm_breedcolour.BreedColourID
LEFT JOIN prm_breedcolourmodifier
ON p.ProfileColourModifierID = prm_breedcolourmodifier.BreedColourModifierID
WHERE p.ProfileName != 'Unknown'
ORDER BY p.ProfileID ASC");
while($row = mysql_fetch_array($sql)) {
$id = $row['ProfileID'];
$name = $row['ProfilePrefix'] . ' ' . $row['ProfileName'];
if ($row['ProfileYearOfDeath'] > 0000) { $age = ($row['ProfileYearOfDeath'] - $row['ProfileYearOfBirth']); }
elseif ($row['ProfileYearOfDeath'] <= 0000) { $age = (date('Y') - $row['ProfileYearOfBirth']); }
$reg = $row['ProfileLocalRegNumber'];
$sire = $row['sirePrefix'] . ' ' . $row['sireName'];
$dam = $row['damPrefix'] . ' ' . $row['damName'];
$colour = $row['BreedColour'];
$gender = $row['BreedGender'];
$owner = $row['owner_fname'] . ' ' . $row['owner_lname'];
$breeder = $row['breeder_fname'] . ' ' . $row['breeder_lname'];
echo '<tr><td>' . $i++ . '</td><td>' . $name . '</td><td>' . $sire . '</td>';
echo '<td>' . $dam . '</td><td>' . $age . '</td><td>' . $colour . '</td><td>' . $gender. '</td>';
echo '<td>' . $owner . '</td><td>' . $breeder. '</td></tr>';
}
echo '</table>';
mysql_close($con);
Use GROUP BY over DISTINCT:
http://msmvps.com/blogs/robfarley/archive/2007/03/24/group-by-v-distinct-group-by-wins.aspx
The problem is going to be in the data - one of the tables that you're joining against has multiple rows on associated to the join key.
I recommend executing the query in stages. Start with the base query (taking out the field list):
SELECT count(*)
FROM profiles AS p
WHERE p.ProfileName != 'Unknown'
And then add the join tables in one at a time until you see the count increase...
SELECT count(*)
FROM profiles AS p
LEFT JOIN profiles AS sire
ON p.ProfileSireReg = sire.ProfileLocalRegNumber
WHERE p.ProfileName != 'Unknown'
You should then be able to see where the duplicate is. If you want to easily see which record is duplicated, you can run this query:
SELECT p.Profile_id, count(*) cnt
FROM profiles AS p
LEFT JOIN profiles AS sire
ON p.ProfileSireReg = sire.ProfileLocalRegNumber
-- (all other joins)
WHERE p.ProfileName != 'Unknown'
GROUP BY p.Profile_id
HAVING count(*) > 1
Then you can look at the details of the duplicated records.