Question

Please consider the following table structure (this is sample data, so please ignore the timestamps being identical):

+---------+---------+------------+-----------+
| list_id | item_id | date_added | is_active |
+---------+---------+------------+-----------+
|    1    |     1   | 1352073600 |     1     |
|    1    |     2   | 1372073600 |     1     |
|    1    |     3   | 1332073600 |     1     |
|    1    |     4   | 1302073600 |     1     |
|    2    |     1   | 1302073600 |     1     |
|    3    |     1   | 1302073600 |     1     |
+---------+---------+------------+-----------+

Our client wishes to show how many lists were created on a certain day. The list is created when the first item is added, but the date is not explicitly stored, only the date the items have been added.

My question is this:

Using MySQL (and PHP for computing the timestamp), how can I return the number of lists that were created on a certain day. Essentially, the logic should be (pseudo-code):

• select total records from tbl_list_items where date_added >= min_age and date_added <= max_age, and record is oldest for this list

It is difficult to explain what I’m looking for, so consider the following actions:

No items added to lists yet (i.e. all lists have 0 items)
User added `item_id = 1` to `list_id = 1` yesterday.
User added `item_id = 2` to `list_id = 1` today.
User added `item_id = 1` to `list_id = 2` yesterday.
User added `item_id = 1` to `list_id = 3` yesterday.
User added `item_id = 2` to `list_id = 2` today.

If we wanted to look how many lists were created yesterday (or rather, how many lists had the first item added yesterday), I would like to return the total number where the item first added = yesterday. Given the action set above, this would return a total of 3 (i.e. list_id = 1, list_id = 2 and list_id = 3).

Answer 1

This returns the list of lists created on a specific day:

SELECT list_id, min(date_added) AS date_creation
FROM tbl_list_items
GROUP BY list_id
HAVING min(date_added) >= @min_age AND min(date_added) <= @max_age

To count them you can count number of results of the query, or use sub-queries like:

SELECT COUNT(*) AS total FROM
(
    SELECT list_id, min(date_added) AS date_creation
    FROM tbl_list_items
    GROUP BY list_id
    HAVING min(date_added) >= @min_age AND min(date_added) <= @max_age
)

However, note that this is totally not-optimized, because you have to GROUP BY the whole table for each query; maybe will be better to add a flag ( creation = 0 | 1 ) when the first item of a list is added.