How to remove duplicate row considering the Arabic Phonetics - php

Question

I have a table of Arabic text. I want to remove duplicate rows. In view of the symbols in Arabic language: َ ِ ُ
My table: vocabulary
+----+----------+--------------------------------+
| id | word | mean |
--------------------------------------------------
| 1 | سِلام | xxx |
--------------------------------------------------
| 2 | سَلام | xxx |
--------------------------------------------------
| 3 | سلام | xxx |
--------------------------------------------------
| 4 | سلام | xxx |
+------------------------------------------------+
Now i want this table:
+----+----------+--------------------------------+
| id | word | mean |
--------------------------------------------------
| 1 | سِلام | xxx |
--------------------------------------------------
| 2 | سَلام | xxx |
--------------------------------------------------
| 3 | سلام | xxx |
+------------------------------------------------+
How can i do that ?!
My Try:
$result = mysql_query( "SELECT * FROM vocabulary where");
while($end = mysql_fetch_assoc($result)){
$word = $end["word"];
$mean = $end["mean"];
$id = $end["id"];
$result2 = mysql_query( "SELECT * FROM vocabulary where word='$word' AND mean='$mean'");
$TotalResults = mysql_num_rows($result2);
if($TotalResults>1){
mysql_query( "DELETE FROM vocabulary WHERE id='$id'");
}
Summary: How can I sensitive MySQL to the Arabic symbols ?

Answer 1

There are multiple ways to achieve this.
1- You can either select your rows from the database, loop through them and save the 'word' title in an array, and in each iteration in the loop, you can check if a similar value is in_array(). If the value exists, then you can save the id in another array and then use these ids to delete from the database.
2- Another way to extract the ids is to use a query similar to the below:
select count(*), id from table group by title
You can then loop through the results and delete the row (using the ids) where count is greater than 1.
The basic concept in both (and other methods) is that you just have to match the strings. Phonetics on letters change the actual string so "سَلام" is not equal to "سلام".
On a side note, there is a great Arabic PHP library you can use for various Arabic related string manipulation: PHP and Arabic Language.
This way will only remove one duplicate.
There are several other ways to do it, and it all depends on the size of the data set you have and if deleting these duplicates is a one time thing or a frequent thing because you will have to keep performance in mind.

Answer 2

I haven't tested it, but this should work:
CREATE TEMPORARY TABLE tmp_keeps
SELECT title, MIN(id) AS keepID
FROM theTable
GROUP BY title
;
DELETE FROM theTable
WHERE (title, id) NOT IN (
SELECT title, keepID
FROM tmp_keeps
)
;
DROP TEMPORARY TABLE tmp_keeps;
It (in the subquery) gets the first id for each title, and then deletes rows that don't meet that condition.
Edit: Revised to avoid SQL error pointed out in comments.
If it is a large table, something along the lines of Adon's answer might be faster.