I’m building a report for a database where I need to determine the number of “first scans” grouping by company, job, and date.
The scan table can contain multiple scans for the same item, however I only want to include the original scan in my COUNT, which can only be identified as being the scan with the earliest date that matches a particular item.
My first attempt at this was:
SELECT
_item_detail.job_id,
_item_group.group_id,
_scan.company_id,
DATE(scan_date_time) as scan_date,
COUNT(1)
FROM _scan
INNER JOIN _item_detail ON _item_detail.company_id = _scan.company_id
AND
_item_detail.serial_number = _scan.serial_number
INNER JOIN _item_group ON _item_group.group_id = _item_detail.group_id
WHERE _item_detail.job_id = '0326FCM' AND _scan.company_id = '152345' AND _item_group.group_id = 13
GROUP BY
_item_detail.job_id,
_item_group.group_id,
_scan.company_id, scan_date -- first_scan_count
HAVING min(scan_date_time);
This is giving me incorrect results, though (about 3x too many). I am assuming it’s because the MIN record is being recalculated for each date, so if the min was found on day 1, it may also be found on day 3 and counted again.
How can I modify my query to achieve the desired results?
Something similar to this should work… I’m not completely sure of how your tables are laid out or how the data relates them together, but this is the general idea: