New and Improved 3.13 Beta. Highlights: Shareable filters and dashboards and lots of other goodies. Any feedback can be raised as JIRA issues in the JIRA project.
Issue Details (XML | Word | Printable)

Key: JRA-6527
Type: Improvement Improvement
Status: Open Open
Priority: Major Major
Assignee: Unassigned
Reporter: Andy Armstrong
Votes: 32
Watchers: 17
Operations

If you were logged in you would be able to see more operations.
JIRA

Allow filters to be built upon other shared filters (combined filters)

Created: 02/May/05 08:50 AM   Updated: 12/Feb/08 05:40 PM
Component/s: Filtering & Indexing
Affects Version/s: 3.1.1
Fix Version/s: None

Time Tracking:
Not Specified

Issue Links:
Duplicate
 
Reference

Participants: Andy Armstrong, Anton Mazkovoi [Atlassian], Gavin Bee, Jeff Turner [Atlassian], John M. Black and John Sumsion
Since last comment: 1 year, 26 weeks, 2 days ago
Labels:


 Description  « Hide
We have around twenty shared filters implemented now, and most of them are subtle variants on a base filter.

e.g. we have a filter called 'current unresolved issues' which is all unresolved issues in the release we are currently working on. We then have:
my current unresolved issues
my team's current unresolved issues
current unresolved issues due this week
current unresolved GUI issues
my current unresolved GUI issues

When we update the definition of 'current unresolved issues' (e.g. we change the 'fix for' field to the new current release) we have to go and update all of the 'sub'-filters.

There are two solutions that would help us here:

  • have a notion of the 'currently active' releases, because then we wouldn't need to keep updating the filter.
  • allow 'my current unresolved issues' to be defined as:
    filter 'current unresolved issues' where user = currentUser


 All   Comments   Work Log   Change History      Sort Order: Ascending order - Click to sort in descending order
Jeff Turner [Atlassian] added a comment - 04/May/05 11:26 PM
Andy,

Thanks for that very clear description. We have been thinking about implementing this recently (filters based on other filters, and/or/not'ed joined). Hopefully it will happen in the next development cycle.


John Sumsion added a comment - 15/May/06 02:33 PM
I already have filters for "Unresolved Issues assigned to me" and "Open Issues submitted by me".

But since I often open a lot of my own issues myself, I really want to see "Open Issues submitted by me (assigned to any team other than mine)". And I thought I could do that by defining an "Open Issues Assigned to my team" filter and defining another filter that is just ("Open Issues submitted by me" minus "Open Issues Assigned to my team").

If the query engine knew that the "Open Issues" were the same set in both queries, that could be optimized out and either cached or only evalutated once.

P.S. Sorry for making this comment twice, I just didn't see this issue until I had already commented on JRA-9629. But the description of this issue is more in line with what I would need, so I wanted to state my needs clearly here.


John Sumsion added a comment - 15/May/06 02:33 PM
Also, it sounds like this issue hasn't made it onto the list for even 3.6, do you have any more refined idea of when you might get to this?

Anton Mazkovoi [Atlassian] added a comment - 15/May/06 10:05 PM
John,

Unfortuantely we do not have an implementation date. The best thing to do is vote for thsi issue.

Thanks,
Anton


Gavin Bee added a comment - 21/Feb/07 10:57 PM
Implementing this issue would assist the many people looking for product suite, sub-project, and sub-component support. Looking through the many issues on this topic, it seems that most people really actually want products.

With this feature, I could create a product version filter (eg Suite 4.0) that combines the results of my "Unresolved App X 4.0", "Unresolved Core 10.2", and "Unresolved SomeOtherSmallProject 4.2.1" filters.


John M. Black added a comment - 27/Feb/07 08:54 AM
Thanks, Andy, for noticing the similarity with JIRA-1560. In that thread we were talking about custom SQL queries to do advance search operations, but now that I read this thread, I realize that a lot of what people want in 1560 can be accomplished by "stacking" queries as described here. (In fact, I think this solution could be adequate enough for 1560's requirements that it could even be considered a duplicate of this.)

The example in 1560 I gave is similar to Andy & John's example – we want a result set of "anything related to Version 123," regardless of whether it's Affects or Fixed-In. This would require an OR search or a "stacking" of two queries.

Pipe dream: How about adding the concept of "runtime variables?" One could define a variable in the individual queries that would only need to be changed once in the whole stack, like:

query 1) where Fixed-In version = var X
query 2) where Affects = var X
Stacked query: (Query 1 || Query 2), and var X = 123

If I then need to change X (which might be very frequent,) I would only need to do it in the stacked query settings, not in both queries 1 & 2. In fact, it could be a runtime selection on the query result screen itself, to save the user from having to go through an Edit path to change it. So I call my query "Related to Specific Version" and run it; and the result screen has a widget to pick the version.