Selecting Grouped Row Count When Using MySQL GROUP BY Column

You may have been using subqueries to get how many rows were grouped when using MySQL’s GROUP BY function. There’s a little trick that helps to get this number without using subqueries, more clean and faster (when the amount of rows get higher).

Let’s assume we have a table contain folks from different cities. Like this:

folk_name   |   city_name
---------------------------
evren       |   ohio
steve       |   ohio
cem         |   weston
don         |   sunderland
ritchie     |   weston
ian         |   hounslow
alican      |   ohio

When we want to get only city names (distinct) from this table, we may group the data by city_name column like this:

SELECT f.* FROM `folks` f GROUP BY f.`city_name`

And the result would be:

folk_name   |   city_name
---------------------------
ian         |   hounslow
evren       |   ohio
don         |   sunderland
cem         |   weston

The expected behaviour of GROUP BY is to select first rows so other folks from same cities do not appear. It also orders by the GROUP BY column (ascending).

Let’s say we want to get how many folks exist in each city by grouping by. First thing comes to mind could be using a subquery like this:

SELECT f1.*, (SELECT COUNT(f2.`folk_name`) FROM `folks` f2 WHERE f2.`city_name` = f1.`city_name`) AS `folk_count` FROM `folks` f1 GROUP BY `city_name`

And the result would be:

folk_name   |   city_name   |   folk_count
-------------------------------------------
ian         |   hounslow    |   1
evren       |   ohio        |   3
don         |   sunderland  |   1
cem         |   weston      |   2

Let’s admit, subqueries are messy. They make the query more complicated to read and also are question marks on performance.

How about selecting number 1 (one) for each row and sum it up on grouping by?

Doing something like this:

SELECT f.*,SUM(1) AS `folk_count` FROM `folks` f GROUP BY f.`city_name`

Would return:

folk_name   |   city_name   |   folk_count
-------------------------------------------
ian         |   hounslow    |   1
evren       |   ohio        |   3
don         |   sunderland  |   1
cem         |   weston      |   2

Nicely done.

Generate DateTime Value in PHP for Using in MySQL Command

When we work with date and/or time values in MySQL, it’s always easy to use built-in MySQL functions such as NOW(), DATE_ADD(), TIMESTAMP() etc… Now imagine you have to pass date/time values to MySQL from PHP variables. First time I tried that, it didn’t work out so easy for me.

Yes, if you simply take one column value and use it again, you have no problem. But here’s an example that you take date and time from MySQL column, add 1 seconds to that value and use it in query again:

//Here's just stuff in order to get a datetime value from MySQL DB.
$query = "SELECT TOP 1 `date` FROM `table`";
$result = mysql_query($query, $db_link);
$date_time_val = "";
if ($result) {
    while($row = mysql_fetch_assoc($result)) {
        //Here we get the value.
        $date_time_val = $row["date"];
    }
}
 
//Now we add one second to this value and use it again.
//First, use it for generating a DateTime variable from PHP library.
$time = new DateTime($date_time_val);
//If you need current time in PHP, you can use this:
//$time = new DateTime(date("Y-m-d H:i:s"));
 
//Add one second,
$time->modify("+1 seconds");
 
//You can not use it as it is
//You need to convert to a string in MySQL datetime format.
$time_string = $time->format("Y-m-d H:i:s");
 
//Now you're free to use in your query
$query_new = "SELECT * FROM `table` WHERE `date` = '".$time_string."' ";

T-SQL: Concatenate Multiple Rows of a Column into One Cell

Simple: You have a column and you want to take all rows as one cell. So that you can do searching, ordering, whatever you like.

Solution: Declare a variable, in this case VARCHAR or NVARCHAR, then SELECT the variable by equaling [itself] plus [column name] plus [your seperation char].

Here’s the code snippet:

DECLARE @concatenated NVARCHAR(1000)
SET @concatenated = ''
SELECT @concatenated = @concatenated + [NAME_COLUMN] + ',' FROM [USERS_TABLE] WHERE [ID_COLUMN] < 100
SELECT @concatenated