Question

I use h2 database.
I have 3 tables: USER, SCHEDULEWORKER and SECURITYACCESS.

USER

userNumber, firstName, lastName

SCHEDULEWORKER

USERNUMBER, STARTDATE, ENDDATE  

SECURITYACCESS

USERNUMBER, DATEACCESS, DOOR

STARTDATE, ENDDATE and DATEACCESS use timestamp datatype.

I would like to get know all user who suppose to work a day and who entered the maindoor after the start date.

So for this example

SCHEDULEWORKER  
USERNUMBER      STARTDATE               ENDDATE  
4141            2012-07-02 07:30:00.0   2012-07-02 16:30:00.0
4123            2012-07-02 07:30:00.0   2012-07-02 16:30:00.0

SECURITYACCESS
USERNUMBER      DATEACCESS      DOOR  
4145    2012-07-02 07:31:56.0   MainDoor
4123    2012-07-02 07:29:56.0   MainDoor

I would like to have

USERNUMBER  STARTDATE              ENDDATE                DATEACCESS        DOOR        late
4141    2012-07-02 07:30:00.0  2012-07-02 16:30:00.0  2012-07-02 07:31:56.0 MainDoor    00:01:56

Do I need to split my field timestamp to be able to do this query?

Answer 1

It will be slow but if you have many SECURITYACESS in a day you have to compare first access that day.
I don’t work with h2 so you have to manage timestamp calculations for late column

select SW.USERNUMBER, STARTDATE, ENDDATE, DATEACCESS, DOOR, (h2 calculations on timestamps) as late
FROM SCHEDULEWORKER AS SW 
  inner join SECURITYACCESS AS SA on SW.USERNUMBER=SA.USERNUMBER
  inner join (SELECT USERNUMBER, MIN(DATEACCESS) as ACCESS
              FROM SECURITYACCESS 
              GROUP BY USERNUMBER, YEAR(DATEACCESS), DAY_OF_YEAR(DATEACCESS)) AS MSA ON SW.USERNUMBER = MSA.USERNUMBER AND MSA.ACCESS = SA.DATEACCESS
WHERE SW.STARTDATE < MSA.ACCESS

If you have more where conditions: date range maybe or usernumber as parameter you should change MSA query for better performance.

I didn’t found timestamp to date conversion or function for better group by clause.