Question

I try to create a MySQL Statement with multiple SQL Statements.

What i trying to do:
I get two tables, project & jiraissue.

Table: project

Columns:
ID  decimal(18,0) PK 
pname   varchar(255) 
URL varchar(255) 
LEAD    varchar(255) 
DESCRIPTION text 
pkey    varchar(255) 
pcounter    decimal(18,0) 
ASSIGNEETYPE    decimal(18,0) 

Table: jiraissue

Columns:
ID  decimal(18,0) PK 
pkey    varchar(255) 
PROJECT decimal(18,0) 
REPORTER    varchar(255) 
ASSIGNEE    varchar(255) 
issuetype   varchar(255) 
SUMMARY varchar(255) 
DESCRIPTION longtext 
ENVIRONMENT longtext 
PRIORITY    varchar(255) 
RESOLUTION  varchar(255) 
issuestatus varchar(255) 
CREATED datetime 
UPDATED datetime 
DUEDATE datetime 
VOTES   decimal(18,0) 
TIMEORIGINALESTIMATE    decimal(18,0) 
TIMEESTIMATE    decimal(18,0) 
TIMESPENT   decimal(18,0) 
WORKFLOW_ID decimal(18,0) 
SECURITY    decimal(18,0) 
FIXFOR  decimal(18,0) 
COMPONENT   decimal(18,0) 

My goal is to get the names of the projects (project.pname) which newest issue is updated before 2012 (jiraissue.UPDATED).

Example:
I get a Project ABC which newest issue was update in 21.11.2012. The other Project XYZ‘s newest issue was updated last in 08.12.2011.
So my SQL Script should give me the name of the second project XYZ but not the first ABC

A working script to find out which issue is the newest with a STATIC project ID comes here:

SELECT 
    pkey
FROM
    jiraissue
WHERE
    UPDATED = (SELECT 
            max(UPDATED)
        FROM
            jiraissue
        WHERE
            PROJECT = 10472)

But how is the script now, when i want the project names of ALL these projects? All of my trys take a lot of processing time and give a undefinied error back…

EDIT:
Now i get the following code:

select 
    p.pname, j.pkey
from
    project p
        inner join
    jiraissue j ON j.ID = (select 
            PROJECT
        from
            jiraissue
        where
            UPDATED = (SELECT 
                    max(UPDATED)
                from
                    jiraissue))
        AND p.ID = j.PROJECT

The Result is but just the first project with the lowest key… How can i browse ALL Projects?

EDIT:

select 
    p.pname, j.pkey, j.UPDATED
from
    project p
        inner join
    jiraissue j
where
    j.ID = (select 
            PROJECT
        from
            jiraissue
        where
            UPDATED = (SELECT 
                    max(UPDATED)
                from
                    jiraissue
                where
                    UPDATED < '2012-01-01 00:00:00'))
        and p.ID = j.PROJECT

It displays all what i want. But just the first project. So i need ALL projects!! How can i select not just 1 project but all projects?

Answer 1

Start by first WHAT do you want. On a per project basis, what is the latest issue that was updated BEFORE 2012. Ignore the project table at the moment..

select
      JI.Project,
      MAX( JI.Updated ) as LastUpdated
   from
      JiraIssue JI
   group by
      JI.Project
   having 
      LastUpdated < '2012-01-01'

Now, this will give you all your primary information that you want. The HAVING clause is applied AFTER the group by to exclude any projects that are 2012 and future. You can’t apply it as the WHERE clause as your sample described you didn’t want anything that had a current (such as your ABC project) and only wanted the XYZ project.

Use this as a basis for getting the rest of your data by joining to the project table and joining AGAIN back to the JiraIssue table by that project/update combination

select
      P.*,
      JI2.*
   from
      ( select
              JI.Project,
              MAX( JI.Updated ) as LastUpdated
           from
              JiraIssue JI
           group by
              JI.Project
           having 
              LastUpdated < '2012-01-01' ) PreQuery

         JOIN Project P
            on PreQuery.Project = P.ID

         JOIN JiraIssue JI2
            on PreQuery.Project = JI2.Project
           AND PreQuery.LastUpdated = JI2.Updated