Need help with turning multiple rows into a single row

Question

Hello.
I've come across a situation that is somewhat beyond my knowledge base. I could use a little help with figuring this out.
My situation:
I am attempting to do some reporting from a JIRA database. What I am doing is getting the dates and times for specific step points of a ticket. This is resulting in many rows per ticket. What I need to do is return one row per ticket with a calculation of time between each step. But one issue is that if a ticket is re-opened, I want to ignore all data beyond the first close date. Also, not all tickets are in a closed state. I am attaching code and a sample list of the results. If I am not quite clear, please ask for information and I will attempt to provide more. The database is 10.2.0.4
select jiraissue.id, pkey, reporter, summary
,changegroup.created change_dt
,dbms_lob.substr(changeitem.newstring,15,1) change_type
,row_number() OVER ( PARTITION BY jiraissue.id ORDER BY changegroup.created ASC ) AS order_row
from jiraissue
,changeitem, changegroup
,(select * from customfieldvalue where customfield = 10591 and stringvalue = 'Support') phaseinfo
where jiraissue.project = 10110
and jiraissue.issuetype = 51
and dbms_lob.substr(changeitem.newstring,15,1) in ('Blocked','Closed','Testing','Open')
and phaseinfo.issue = jiraissue.id
and changeitem.groupid = changegroup.id
and changegroup.issueid = jiraissue.id
order by jiraissue.id,change_dt
Results:
1     21191     QCS-91     Error running the Earliest-deadlines flight interface request/response message     2008-07-16 9:30:38 AM     Open     1
2     21191     QCS-91     Error running the Earliest-deadlines flight interface request/response message     2008-07-16 11:37:02 AM     Testing     2
3     21191     QCS-91     Error running the Earliest-deadlines flight interface request/response message     2010-06-08 9:14:52 AM     Closed     3
4     21191     QCS-91     Error running the Earliest-deadlines flight interface request/response message     2010-09-02 11:29:37 AM     Open     4
5     21191     QCS-91     Error running the Earliest-deadlines flight interface request/response message     2010-09-02 11:29:42 AM     Open     5
6     21191     QCS-91     Error running the Earliest-deadlines flight interface request/response message     2010-09-02 11:29:50 AM     Testing     6
7     21191     QCS-91     Error running the Earliest-deadlines flight interface request/response message     2010-09-02 11:29:53 AM     Closed     7
8     23234     QCS-208     System Baseline - OK button does not show up in the Defer Faults page for the System Engineer role      2008-10-03 10:26:21 AM     Open     1
9     23234     QCS-208     System Baseline - OK button does not show up in the Defer Faults page for the System Engineer role      2008-11-17 9:39:39 AM     Testing     2
10     23234     QCS-208     System Baseline - OK button does not show up in the Defer Faults page for the System Engineer role      2011-02-02 6:18:02 AM     Closed     3
11     23977     QCS-311     Tally Sheet - Reason Not Done fails to provide reason for unassigned tasks     2008-09-29 2:44:54 PM     Open     1
12     23977     QCS-311     Tally Sheet - Reason Not Done fails to provide reason for unassigned tasks     2010-05-29 4:47:37 PM     Blocked     2
13     23977     QCS-311     Tally Sheet - Reason Not Done fails to provide reason for unassigned tasks     2011-02-02 6:14:57 AM     Open     3
14     23977     QCS-311     Tally Sheet - Reason Not Done fails to provide reason for unassigned tasks     2011-02-02 6:15:32 AM     Testing     4
15     23977     QCS-311     Tally Sheet - Reason Not Done fails to provide reason for unassigned tasks     2011-02-02 6:15:47 AM     Closed     5

Answer 1

Hi,
Welcome to the forum!
StblJmpr wrote:
... I am attempting to do some reporting from a JIRA database. What is a JIRA database?
I am attaching code and a sample list of the results. If I am not quite clear, please ask for information and I will attempt to provide more. Whenever you have a question, please post a little sample data (CREATE TABLE and INSERT statements, relevant columns only) for all tables involved, and the results you want from that data.
Simplify the problem as much as possible. For example, if the part you don't know how to do only involves 2 tables, then jsut post a question involving those 2 tables. So you might just post this much data:
CREATE TABLE     changegroup
(      issueid          NUMBER
,      created          DATE
,      id          NUMBER
INSERT INTO changegroup (issueid, created, id) VALUES (21191,  TO_DATE ('2008-07-16 09:30:38 AM', 'YYYY-MM-DD HH:MI:SS AM'),  10);
INSERT INTO changegroup (issueid, created, id) VALUES (21191,  TO_DATE ('2008-07-16 11:37:02 AM', 'YYYY-MM-DD HH:MI:SS AM'),  20);
INSERT INTO changegroup (issueid, created, id) VALUES (21191,  TO_DATE ('2010-06-08 09:14:52 AM', 'YYYY-MM-DD HH:MI:SS AM'),  90);
INSERT INTO changegroup (issueid, created, id) VALUES (21191,  TO_DATE ('2010-09-02 11:29:37 AM', 'YYYY-MM-DD HH:MI:SS AM'),  10);
INSERT INTO changegroup (issueid, created, id) VALUES (21191,  TO_DATE ('2010-09-02 11:29:42 AM', 'YYYY-MM-DD HH:MI:SS AM'),  10);
INSERT INTO changegroup (issueid, created, id) VALUES (21191,  TO_DATE ('2010-09-02 11:29:50 AM', 'YYYY-MM-DD HH:MI:SS AM'),  20);
INSERT INTO changegroup (issueid, created, id) VALUES (21191,  TO_DATE ('2010-09-02 11:29:53 AM', 'YYYY-MM-DD HH:MI:SS AM'),  90);
INSERT INTO changegroup (issueid, created, id) VALUES (23234,  TO_DATE ('2008-10-03 10:26:21 AM', 'YYYY-MM-DD HH:MI:SS AM'),  10);
INSERT INTO changegroup (issueid, created, id) VALUES (23234,  TO_DATE ('2008-11-17 09:39:39 AM', 'YYYY-MM-DD HH:MI:SS AM'),  20);
INSERT INTO changegroup (issueid, created, id) VALUES (23234,  TO_DATE ('2011-02-02 06:18:02 AM', 'YYYY-MM-DD HH:MI:SS AM'),  90);
INSERT INTO changegroup (issueid, created, id) VALUES (23977,  TO_DATE ('2008-09-29 02:44:54 PM', 'YYYY-MM-DD HH:MI:SS AM'),  10);
INSERT INTO changegroup (issueid, created, id) VALUES (23977,  TO_DATE ('2010-05-29 04:47:37 PM', 'YYYY-MM-DD HH:MI:SS AM'),  30);
INSERT INTO changegroup (issueid, created, id) VALUES (23977,  TO_DATE ('2011-02-02 06:14:57 AM', 'YYYY-MM-DD HH:MI:SS AM'),  10);
INSERT INTO changegroup (issueid, created, id) VALUES (23977,  TO_DATE ('2011-02-02 06:15:32 AM', 'YYYY-MM-DD HH:MI:SS AM'),  20);
INSERT INTO changegroup (issueid, created, id) VALUES (23977,  TO_DATE ('2011-02-02 06:15:47 AM', 'YYYY-MM-DD HH:MI:SS AM'),  90);
CREATE TABLE     changeitem
(      groupid          NUMBER
,      newstring     VARCHAR2 (10)
INSERT INTO changeitem (groupid, newstring) VALUES (10, 'Open');
INSERT INTO changeitem (groupid, newstring) VALUES (20, 'Testing');
INSERT INTO changeitem (groupid, newstring) VALUES (30, 'Blocked');
INSERT INTO changeitem (groupid, newstring) VALUES (90, 'Closed');Then post the results you want to get from that data, like this:
ISSUEID HISTORY
  21191 Open (0) >> Testing (692) >> Closed
  23234 Open (45) >> Testing (807) >> Closed
  23977 Open (607) >> Blocked (249) >> Open (0) >> Testing (0) >> ClosedExplain how you get those results from that data. For example:
"The output contains one row per issueid. The HISTORY coloumn shows the different states that the issue went through, in order by created, starting with the earliest one and continuing up until the first 'Closed' state, if there is one. Take the first row, issueid=21191, for example. It started as 'Open' on July 16, 2008, then, on the same day (that is, 0 days later) changed to 'Testing', and then, on June 8, 2010, (692 days later), it became 'Closed'. That same issue opened again later, on September 2, 2010, but I don't want to see any activity after the first 'Closed'."
The database is 10.2.0.4That's very important. Always post your version, like you did.
Here's one way to get those results from that data:
WITH     got_order_row     AS
     SELECT     cg.issueid
     ,     LEAD (cg.created) OVER ( PARTITION BY  cg.issueid
                                      ORDER BY      cg.created
              - cg.created            AS days_in_stage
     ,       ROW_NUMBER ()     OVER ( PARTITION BY  cg.issueid
                                      ORDER BY      cg.created
                           )    AS order_row
     ,     ci.newstring                     AS change_type
     FROM    changegroup     cg
     JOIN     changeitem     ci  ON   cg.id     = ci.groupid
     WHERE     ci.newstring     IN ( 'Blocked'
                       , 'Closed'
                       , 'Testing'
                       , 'Open'
--     AND     ...          -- any other filtering goes here
SELECT       issueid
,       SUBSTR ( SYS_CONNECT_BY_PATH ( change_type || CASE
                                                         WHEN  CONNECT_BY_ISLEAF = 0
                                       THEN  ' ('
                                          || ROUND (days_in_stage)
                                          || ')'
                                                     END
                                , ' >> '
           , 5
           )     AS history
FROM       got_order_row
WHERE       CONNECT_BY_ISLEAF     = 1
START WITH     order_row          = 1
CONNECT BY     order_row          = PRIOR order_row + 1
     AND     issueid               = PRIOR issueid
     AND     PRIOR change_type     != 'Closed'
ORDER BY  issueid
;Combining data from several rows into one big delimited VARCHAR2 column on one row is call String Aggregation .
I hope this answers your question, but I guessed at so many things, I won't be surprised if it doesn't. If that's the case, point out where this is wrong, post what the results should be in those places, and explain how you get those results. Post new data, if necessary.