Scenario
I have a MySQL database with 10.000 rows. Setup of the database:
ID UniqueKey Name Url Score ItemValue
1 5Zvr3 Google google.com 13 X
2 46cfG Radio radio.com -20 X
3 2fg64 Yahoo yahoo.com 5 X
.... etc etc etc
As you can see, each item has a score. The score is constantly changing. Google may have a score of 13 now, but tomorrow it may be 80, or -50.
What I want:
I want to create a system that creates a hierarchy in my current database, based on the score of the items. Right now I'm thinking about percentile ranks, meaning that the highest scoring items will be close to 100%, and the lowest scoring items will be close to 0%. For this I created some code that will try to achieve what is shown here: http://www.psychstat.missouristate.edu/introbook/sbk14m.htm
This is my code:
$sql = "SELECT * FROM database order by Score";
$result = $conn->query($sql);
$count = 0;
while ($row = $result->fetch_assoc()) {
$woow = $row['Score'];
$sql = "SELECT * FROM database WHERE Score = $woow";
$resultnew = $conn->query($sql);
$somanythesame = $resultnew->num_rows;
$itemPercentile = ( ($count/$result->num_rows + 0.5*$somanythesame/$result->num_rows) * 100 );
$rowID = $row['ID'];
$sql2 = "UPDATE database SET itemValue = $itemPercentile WHERE ID = $rowID";
$conn->query($sql2);
$count++;
}
This works, but for one problem it does not: There are many items in my database, many with the same score. To illustrate my problem, here is a very simple 10-row database with only the Scores:
Scores
-10
0
0
0
10
20
20
30
40
50
The problem with my code is that it doesn't give the same percentile for the items with the same Score, because it takes in account all previous rows for the calculation, including the ones with the same Score.
So, for the 2nd, 3rd and 4th item with a Score of 0, it should be like this: (1/10 + 0.5*1/10) * 100. Problem is, that for the 3rd item it will do (2/10 + 0.5*1/10) * 100 and the 4th item it will do (3/10 + 0.5*1/10) * 100.
Then, for the 5th item with a score of 10, it should do (4/10 + 0.5*1/10) * 100. This is going well; only not for the items with te same score.
I'm not sure if I explained this well, I find it hard to put my problem in the right words. If you have any questions, let me know! Thank you for your time :)
You need to maintain an "identical count" ($icount) variable that tracks the number of items with an identical score and a "current score" ($score) that tracks the current score.
$icount = 0;
$score = null;
Increment $icount instead of $count when $woow == $score (identical value check). Otherwise, add it to your $count and increment, and then reset the $icount value to 0.
if ($woow == $score) {
$icount++;
} else {
$count += $icount + 1;
$icount = 0;
}
Finally, set your $score value to the latest $woow for testing in the next iteration of the loop:
$score = $woow;
This will allow items with the same Score to have the same $count value, while incrementing an additional $icount times when a new $score is found.
Your final code will look like this:
$sql = "SELECT * FROM database order by Score";
$result = $conn->query($sql);
$count = 0;
$icount = 0;
$score = null;
while ($row = $result->fetch_assoc()) {
$woow = $row['Score'];
$sql = "SELECT * FROM database WHERE Score = $woow";
$resultnew = $conn->query($sql);
$somanythesame = $resultnew->num_rows;
$itemPercentile = ( ($count/$result->num_rows + 0.5*$somanythesame/$result->num_rows) * 100 );
$rowID = $row['ID'];
$sql2 = "UPDATE database SET itemValue = $itemPercentile WHERE ID = $rowID";
$conn->query($sql2);
if ($woow == $score) {
$icount++;
} else {
$count += $icount + 1;
$icount = 0;
}
$score = $woow;
}
You can change $sql query:
$sql = "SELECT *,count(*) FROM database group by Score order by Score";
In this case, you fetch score with counts and no more select needed in the while loop.
Even you can select Percentile in MySQL query:
Select t2.* , #fb as N , ((t2.fb1 + 0.5 * t2.fw)/#fb*100) as percentile from (
Select t1.* , (#fb := #fb + t1.fw) as fb1 from (
Select score,count(*) as fw From tablename group by score order by score ASC
) as t1
) as t2
I think this query returns most of columns which you may needs to check results.
Related
I have a table called feedbacks that looks like this:
id user type
1 JOhnT Positive
2 JOhnT Negative
3 Sarah Positive
4 JOhnT Positive
5 JOhnT Neutral
....................
I need to get the percentage of POSITIVE feedback for each user using PHP.
I tried something like this which i know it is wrong:
$sql = "SELECT type, count(*),
concat(round(( count(*)/(SELECT count(*) FROM feedbacks WHERE user='JOhnT' AND type='POSITIVE') * 100 ),2),'%') AS percentage
FROM feedback
WHERE user='$email' AND type='positive' GROUP BY type";
$query = mysqli_query($db_conx, $sqlJ);
echo $sql;
Could someone please advice on how to achieve this?
EDIT:
Based on the comments, this is what i have so far:
$sql = "SELECT * FROM feedbacks WHERE user='JOhnT' AND type='POSITIVE'";
$query = mysqli_query($db_conx, $sql) or die(mysqli_error($db_conx));
$productCount = mysqli_num_rows($query);
if ($productCount > 0) {
while($row = mysqli_fetch_array($query, MYSQLI_ASSOC)){
///do I need to calculate the percentage here?//
//if so, how?////
}
}
I ended up creating an array and push each $row in my whole loop into the array and then calculate the percentage like this:
$c = count($values);
$array = $values;
function array_avg($array, $round=1){
$num = count($array);
return array_map(
function($val) use ($num,$round){
return array('count'=>$val,'avg'=>round($val/$num*100, $round));
},
array_count_values($array));
}
$rating = 0;
if($c > 0){
$avgs = array_avg($array);
$rating = $avgs["positive"]["avg"];
}
This works fine for now.
You can do this a conditional average. In MySQL, you could phrase this as:
select user, avg(type = 'Positive') positive_ratio
from feedback
group by user
This gives you, for each user, a value between 0 and 1 that represents the ratio of positive types. You can multiply that by 100 if you want a percentage.
If you want this information for a single user, you can filter with a where clause (and there is no need to group by):
select user, avg(type = 'Positive') positive_ratio
from feedback
where user = 'JOhnT'
Side note: user is a MySQL keyword, hence not a good choice for a column name.
Example column in sql:
id|viewed
1|1000
35|500
47|1200
79|700
84|300
I use "SELECT * FROM table ORDER BY viewed desc", the result I get:
47|1200
1|1000
79|700
35|500
84|300
But I want to add "number order" by Top id ( like ranking ) in PHP, example text:
ID 47 is 1
ID 1 is 2
ID 79 is 3
ID 35 is 4
ID 84 is 5
So what I need to add these top with PHP guys, I think will use $i++ to count number or something like this in php ?
Increment a variable during the loop that prints the results.
$i = 1;
while ($row = $stmt->fetch()) {
echo "ID {$row['id']} is $i<br>";
$i++;
}
You can do it from the sql query itself :
May be you should try this:
set #row_num = 0;
SELECT #row_num := #row_num + 1 AS number,id,viewed FROM `table` ORDER BY viewed desc
Something like this?
$query = $db->query("SELECT * FROM table ORDER BY viewed desc");
$rows = $query->fetchAll(PDO:FETCH_ASSOC);
count = count($rows);
echo '<ul>';
for ($x = 0; $x < count; $x++) {
echo '<li>ID ' + $rows[$x]['id'] + ' is ' + $x + '.</li>'
}
echo '</ul>';
I have two queries, one which will get a list of IDs from the main table, and another which will get all records in another table that relate to that ID.
I'd like to split it into pages of results, but my google searches are only coming up with people who want to have a certain number of results per column value, not who want to limit the overall count by it.
The main table has a column that contains the number of records, so actually reading the other table isn't needed. The limit would be used as a minimum value, so if there are still records left in the current group after the limit, it'll continue displaying them. That'd be somehow calculated as part of the offset so it can start in the correct place.
Here's an example of what I mean:
Table 1:
ID | Records
1 | 2
2 | 3
3 | 28
4 | 7
...
Table 2 (contents don't need to be known for this question):
ID | GroupID | Value
1 | 1 | x
2 | 1 | x
3 | 2 | x
4 | 2 | x
5 | 2 | x
6 | 3 | x
...
If the limit was given as 3 for example, both 1 and 2 should display on the first page, since just 1 by itself is under the limit. The next page will then start on 3, and that'll take up the entire page.
I could just manually count up using PHP until I reach the limit, though it might end up going slow if there were a lot of pages (I've no idea if mysql would be any better in that regard though). Here's a quick example of how that'd work to get the offset:
$page = 2;
$limit = 40;
$count = 0;
$current_page = 1;
$query = 'SELECT ID, Records FROM table1';
$stmt = $conn->prepare($query);
$stmt->execute();
while($row = $stmt->fetch(PDO::FETCH_ASSOC)){
$count += $row['Records'];
if($count > $limit){
$current_page ++;
$count = 0;
if($current_page == $page){
$start_id = $row['ID'];
break;
}
}
}
Updated:
A "custom" Pagination is what you're looking for. So if you're planning to hop from page to page, you can't use hardcoded $page and $current_page values. Those values should be generated dynamically once you're on a particular page. In fact, you should have the ID column value in the query part of the URL so that the pagination links could satisfy your business logic.
Assuming the fact that your ID column value starts from 1, your code should be like this:
$id = isset($_GET['id']) && is_numeric($_GET['id']) ? $_GET['id'] : 1;
$limit = 40;
// Display query results based on particular ID number
$query = 'SELECT ID, Records FROM table1 WHERE ID >= ' . $id;
$stmt = $conn->prepare($query);
$stmt->execute();
$rowCount = $stmt->rowCount();
if($rowCount){
$total = 0;
while($row = $stmt->fetch(PDO::FETCH_ASSOC)){
if($total <= $limit){
$total += $row['Records'];
// Display $row details here
}else{
// Get the next $id value
$id = $row['ID'];
break;
}
}
}
// Display relevant ID links
$query = 'SELECT * FROM table1';
$stmt = $conn->prepare($query);
$stmt->execute();
$idArr = array();
$total = 0;
while($row = $stmt->fetch(PDO::FETCH_ASSOC)){
if($total == 0){
$idArr[] = $row['ID'];
}
if($total <= $limit){
$total += $row['Records'];
}else{
$total = 0;
}
}
foreach($idArr as $idValue){
if($idValue == $id){
echo '<a>'.$idValue.'</a> ';
}else{
echo ''.$idValue.' ';
}
}
Sidenote: If the ID column value of your table doesn't start from 1, then use a separate query to get the first ID value and update the following statement,
$id = isset($_GET['id']) && is_numeric($_GET['id']) ? $_GET['id'] : <FIRST_ID_VALUE>;
If you only need to create the "Next" button, you could try this way:
select t1.ID, t1.Records
from table1 t1
left join table1 t2
on t2.ID < t1.ID
and t2.ID > :last_selected_id_1
where t1.ID > :last_selected_id_2
group by t1.ID, t1.Records
having coalesce(sum(t2.Records), 0) < :limit
order by t1.ID
:last_selected_id_x is the last ID from the current page or 0 for the first page.
http://rextester.com/FJRUV28068
You can use MySQL session variables to create page links:
select page, min(ID) as min_id, max(ID) as max_id
from (
select ID
, #page := case when #sum = 0 then #page + 1 else #page end as page
, #sum := case when (#sum + Records) >= :limit
then 0
else #sum + Records
end as sum
from table1
cross join (select #sum := 0, #page := 0) initvars
order by ID
) sub
group by page
The result would look like:
page | min_id | max_id
1 | 1 | 2
2 | 3 | 3
3 | 4 | 4
http://rextester.com/LQVJWP18655
Note that it is officially (as of documentation) not recomended to use the session variables like that (read and write in one statement). Feature versions may break your code.
Here's the result I ended up with, I ended up basing it off Rajdeeps answer but tweaked so that it'll allow sorting.
$query = 'SELECT ID, Records FROM table1 ORDER BY something';
$stmt = $conn->prepare($query);
$stmt->execute();
$id_array = array(); // For storing each ID
$current_ids = array(); // For storing the current ID pages
$initial_ids = array(); // For storing the first page, in case the page is too high
$count = 0;
$current_page = 1;
while($row = $stmt->fetch(PDO::FETCH_ASSOC)){
$already_reset = false;
// Move the session to the next page if it's more over the limit than under
// For example, if a limit of 10 has 6 records already, and the next ID is 9 records, then start a new page
if($id_array and $count + $row['Records'] / 2 > $max_links){
array_push($id_array, $last_id);
$current_page ++;
$count = 0;
$already_reset = true; // Set a marker so that if this ID is larger than the limit then don't start a 2nd new page
}
// Backup the first results in case page is too high and needs to be reset to 1
if($current_page == 1){
array_push($initial_ids, $row['ID']);
}
$count += $row['Records'];
// The values that appear here are the results for the selected page
// They may not be in sequence so store the individual IDs
if($_GET['page'] == $current_page){
array_push($current_ids, $row['ID']);
}
// Start a new page if over the limit
if($count > $max_links and !$already_reset){
$current_page ++;
$count = 0;
array_push($id_array, $row['ID']);
}
$last_id = $row['ID'];
}
array_push($id_array, $last_id);
$total_pages = count($id_array);
// If page is invalid revert to default
if(!$current_ids){
$current_ids = $initial_ids;
$_GET['page'] = 1;
}
$_GET['page'] = max(1, min($total_pages, intval($_GET['page'])));
$current_ids is an array of the IDs for the current page, and $total_pages is pretty self explanatory.
In my project I am fetching data from the database using a while loop where each data is fetched by a customer id as $cid. Now all the data is calculated and finally put into a variable $cashflow. Now suppose there are 20 persons so the table will display 20 records where each data is counted for each person / customer. Now what I want to do is count the total of all 20 records and echo it seperately.
For eg.
Cust.1 : 500 (calculated from $cashflow)
Cust. 2 : 1000
Cust.3 : 800 ..... and so on.
Now these data are well echoed in my table. Now what I want is to add 500+1000+800 and echo 2300 as total separately. But I am not being able to figure out how to do it in my code. Should I take a variable in loop as a ++ counter? or something else? Please help.
Code:
$cash = "SELECT SUM(paid) FROM hbil WHERE cid = '".$cid."' AND method = 'cash'";
$cash_qur = mysql_query($cash) or die(mysql_error());
$cash_neg = "SELECT SUM(price) FROM hbil WHERE cid = '".$cid."' AND method = 'cash' AND price < 0";
$cash_qur_neg = mysql_query($cash_neg) or die(mysql_error());
while($cash_fetch = mysql_fetch_array($cash_qur))
{
$cash_fet = mysql_fetch_array($cash_qur_neg);
$cashpay = $cash_fetch['SUM(paid)'];
$cashneg = $cash_fet['SUM(price)'];
$cash_total = $cashpay-$cashneg;
$cashflow = number_format($cash_total,2,'.',',');
}
$total = 0;
while (...) {
//...
$total += $cash_total;
}
echo $total;
Will do the trick
I'm really stuck and have been searching for a while, but with no success.
Anyway, I have this formula known as the Bayesian Estimate: (WR) = (v ÷ (v+m)) × R + (m ÷ (v+m)) × C
And I have three database tables: persons, reviews, ratings.
The persons table is quite basic, but for the sake of this question, it only has one field: ID
The reviews table has id, personID, description where personID is the ID of the person.
The ratings table has id, personID, reviewID, ratingX, ratingY, ratingZ where person is the ID of the person and reviewID is the ID of the review. ratingX/Y/Z are three different ratings for the person, and in my page shows the average of the three numbers.
THE FORM THAT LISTS THEM, however, sorts them by the Bayesian Estimate formula. I do not know how to do this and it seems beyond me, since you cannot ORDER BY $bayesian_formula or anything like that. The script looks something like this:
<?php
$result = $db->query("SELECT * FROM persons");
$m =; //SQL to get average number of reviews of all persons
$c =; //SQL to get average rating of all persons
while( $row = $result->fetch_array() ){
$r =; //SQL to get average ratings of person
$v =; //SQL to get total number of reviews of person
$formula = ($v / ($v+$m)) * $r + ($m / ($v+$m)) * $c;
$result2 = $db->query("SELECT * FROM ratings ORDER BY $formula");
while( $row2 = $result2->fetch_array() ){
$result3 = $db->query("SELECT * FROM persons WHERE id='$row2[person]'");
$row3 = $result3->fetch_array();
echo $row2['description']."<br> rating: ".round( ($row['ratingX'] + $row['ratingY'] + $row['ratingZ']) / 3 );
}
}
?>
$formula loops to the database result's weighted rating each iteration.
Obviously that isn't correct. How would I make this work? Would I have to revise my entire script? The actual one is much longer and detailed.
edit:
sqls are:
$c_query=$db->query("SELECT ((ratingX + ratingY + ratingZ) / 3) as avg_rate FROM ratings");
$c_ = $c_query->fetch_array();
$c = $c_['avg_rate'];
$m_query=$db->query("SELECT COUNT(id) AS count FROM ratings");
$m_ = $m_query->fetch_array();
$m = $m_['count'];
$v_query=$db->query("SELECT COUNT(id) AS count FROM ratings WHERE person='$row[id]'");
$v_ = $v_query->fetch_array();
$v = $v_['count'];
$r_query=$db->query("SELECT ((ratingX + ratingY + ratingZ) / 3) as avg_rate FROM ratings WHERE person='$row[id]'");
$r_ = $r_query->fetch_array();
$r = $r_['avg_rate'];
You can use CREATE PROCEDURE