Question

I think most people had serious problems with inserting a value/data into a database (mysql). When I’m inserting a value into a database, I assign an unique id (INT) for that line. When I query the database, easily I can read/modify/delete that line.

With function for() (in php) I easily can read values/data from the database. The problem occurs when I delete a line (in the middle for example).

E.g:

DB:

ID | column1 | column2 | ... | columnN
--------------------------------------
1 | abcdefgh | asdasda | ... | asdasdN
2 | asdasddd | asdasda | ... | asdasdN
...
N | asdewfew | asddsad | ... | asddsaN

php:

for($i = 0; $i <= $n; $i++){
   $sql = mysql_query("SELECT * FROM db WHERE ID = '$i' ");
   //Code;
}

*$n = last column value from ID

---

Am I need to reorganize the entire database to have a correct “flow” (1, 2, 3, .. n)? Or am I need to UPDATE the each cell?

Answer 1

What you’re doing here is unnecessary thanks to AUTO_INCREMENT in mysql. Run this command from PhpMyAdmin (or another DB management system):

ALTER TABLE db MODIFY COLUMN ID INT NOT NULL AUTO_INCREMENT;

Now when insert a row into db mysql will assign the ID for you:

INSERT INTO db (id, column1, column2) VALUES(NULL, 'abc', 'def');

SELECT * FROM db;:

+--+-------+-------+
|id|column1|column2|
+--+-------+-------+
|1 |old    |oldrow |
+--+-------+-------+
|2 |abc    |def    | <--- Your newly inserted row, with unique ID
+--+-------+-------+

If you delete a row, it is true that there will be an inconsistency in the order of the ID’s, but this is ok. IDs are not intended to denote the numeric position of a row in a table. They are intended to uniquely identify each row so that you can perform actions on it and reference it from other tables with foreign keys.

---

Also if you need to grab a group of ID’s (stored in an array, for example), it is much more efficient to perform one query with an IN statement.

$ids = array(1,2,3,4,5,6);
$in = implode(',', $ids);
mysql_query('SELECT * FROM db WHERE id IN ('.$in.')');

However, if you want all rows just use:

SELECT * FROM dbs;

But be weary of bobby tables.