Grouper Users - Open Discussion List

[Chris Hyzer <>]

I can easily add a global flag.  Here is my (amended) current plan which I am 
halfway done implementing, and this is my last development item for 1.4 :)

0. If a group is not mentioned in a loader SQL_TABLE, and it doesn’t match 
the name pattern, or the name pattern does not exist, then it will not change
1. If a group is not mentioned in a loader SQL_TABLE, but it does match the 
table pattern, then its members will be removed (no matter what)
2. If a group is not mentioned in a loader SQL_TABLE, but it does match the 
table pattern, and the global switch says no, it wont be deleted
3. If a group is not mentioned in a loader SQL_TABLE, but it does match the 
table pattern, and the global switch say ok, and the group is used somewhere 
(member or composite of another group), then it will not be deleted
4. If a group is not mentioned in a loader SQL_TABLE, but it does match the 
table pattern, and the global switch say ok, and the group is not used 
anywhere, but it is include/exclude, and there are members in the include or 
exclude list, then everything will be deleted except that include or exclude 
list
5. If a group is not mentioned in a loader SQL_TABLE, but it does match the 
table pattern, and the global switch say ok, and the group is not used 
anywhere, but it is include/exclude, and there are no members in the include 
or exclude list, then all related groups will be deleted (include/exclude, 
requiredGroups, etc).

Also, this is leaves a blank stem, that blank stem will be deleted too...

Sound good?

Chris

> -----Original Message-----
> From: Loris Bennett 
> [mailto:]
> Sent: Tuesday, November 11, 2008 9:39 AM
> To: Chris Hyzer
> Cc: Grouper Users Mailing List
> Subject: RE: [grouper-users] Grouper-WS: Strategies for moving
> groups/stems?
>
> Would it be possible to make the removal configurable? It has occurred
> to me that there could be a situation in which a group is only
> temporarily empty and is still needed. Implementing a mechanism for
> defining which groups should be deleted and which shouldn't might be a
> bit complicated, but it would be handy if the deletion could at least
> be
> turned on and off globally.
>
> Thanks
>
> Loris
>
> On Mon, 2008-11-10 at 09:43 -0500, Chris Hyzer wrote:
> > I was planning to do this (probably today) for 1.4...
> >
> > > -----Original Message-----
> > > From: Loris Bennett 
> > > [mailto:]
> > > Sent: Monday, November 10, 2008 4:38 AM
> > > To: Chris Hyzer; Grouper Users Mailing List
> > > Subject: RE: [grouper-users] Grouper-WS: Strategies for moving
> > > groups/stems?
> > >
> > > Could that be included in 1.5?
> > >
> > > Until it is available I shall implement a solution using getGroups.
> > >
> > > Thanks
> > >
> > > Loris
> > >
> > > On Fri, 2008-11-07 at 09:46 -0500, Chris Hyzer wrote:
> > > > I could do that in the loader... if a group is not used anywhere,
> > > remove it.
> > > > Also, there is a web service to list the groups an entity is the
> > > member of: getGroups:
> > > >
> > > >
> https://wiki.internet2.edu/confluence/display/GrouperWG/Get+Groups
> > > >
> > > > Chris
> > > >
> > > > > -----Original Message-----
> > > > > From: Dr. Loris Bennett 
> > > > > [mailto:]
> > > > > Sent: Friday, November 07, 2008 3:25 AM
> > > > > To: Chris Hyzer
> > > > > Subject: RE: [grouper-users] Grouper-WS: Strategies for moving
> > > > > groups/stems?
> > > > >
> > > > > Hi Chris,
> > > > >
> > > > > I have been talking to some colleagues about our needs for
> removing
> > > > > members from groups which no longer appear in the query.
> > > > >
> > > > > Am I correct in thinking that a missing group would have all
> its
> > > > > members
> > > > > removed but the group itself would not be deleted?
> > > > >
> > > > > In this case we would then need to implement some mechanism to
> > > check
> > > > > whether the empty group is used in any other groups before we
> could
> > > > > actually remove it.
> > > > >
> > > > > Looking briefly at the web service I couldn't see a method
> which
> > > would
> > > > > give me the groups an entity is a member of, so I assume I
> would
> > > have
> > > > > to
> > > > > go through all the groups and check if the empty group is used
> > > > > anywhere.
> > > > > Or am I missing something?
> > > > >
> > > > > Regards
> > > > >
> > > > > Loris
> > > > >
> > > > > PS I noticed the release notes for Grouper 1.4 were updated
> > > recently.
> > > > > Is
> > > > > the actual release date imminent?
> > > > >
> > > > >
> > > > >
> > > > > On Fri, 2008-10-17 at 05:07 -0400, Chris Hyzer wrote:
> > > > > > Ok, so the two requested changes to the loader are:
> > > > > >
> > > > > > 1. Dry run mode where it just print out what it will do, but
> > > doesn't
> > > > > actually do it
> > > > > > 2. Attribute which has the SQL "like" pattern of the group
> names
> > > that
> > > > > loader job is responsible for, so it can remove memberships
> from
> > > groups
> > > > > not listed in the query
> > > > > >
> > > > > > Are you also interested in the idea of auto-creating the
> > > composite
> > > > > groups, perhaps with a built-in group type?  And the loader
> having
> > > a
> > > > > way to associate a created group with that type, which would
> auto-
> > > > > create those composites?
> > > > > >
> > > > > > As far as timeline, these changes should be there in the next
> > > release
> > > > > which is the end of the month.  And if not, then shortly
> > > thereafter.
> > > > > >
> > > > > > Also we discussed the group detail being created by web
> service,
> > > > > which you said was not as important.
> > > > > >
> > > > > > Regards,
> > > > > > Chris
> > > > > >
> > > > > > > -----Original Message-----
> > > > > > > From: Loris Bennett 
> > > > > > > [mailto:-
> berlin.de]
> > > > > > > Sent: Friday, October 17, 2008 4:40 AM
> > > > > > > To: Chris Hyzer
> > > > > > > Subject: RE: [grouper-users] Grouper-WS: Strategies for
> moving
> > > > > > > groups/stems?
> > > > > > >
> > > > > > > On Thu, 2008-10-16 at 11:59 -0400, Chris Hyzer wrote:
> > > > > > > > >
> > > > > > > > > Deleting empty groups my hand would be OK for us. We
> are
> > > > > already a
> > > > > > > > > little chicken about the idea of updating our basic org
> > > groups
> > > > > 100%
> > > > > > > > > automatically and hoping that nothing breaks in the
> > > composite
> > > > > > > groups.
> > > > > > > > > Ideally we would like to generate some sort of summary
> of
> > > the
> > > > > > > changes
> > > > > > > > > before these are actually propagated. This would give
> us an
> > > > > > > overview of
> > > > > > > > > the changes occurring and enable us to anticipate the
> > > effects
> > > > > on
> > > > > > > > > derived
> > > > > > > > > groups.
> > > > > > > >
> > > > > > > > Are you saying that all grouper loader changes should
> > > generate a
> > > > > > > summary, or only when you delete groups?
> > > > > > >
> > > > > > > At least initially, we would like a summary of all group
> > > additions
> > > > > and
> > > > > > > deletions in our org data. However, I see this as
> essentially
> > > > > > > independent of grouper-loader, although I suppose it could
> be
> > > built
> > > > > > > into
> > > > > > > grouper-loader as a dry run mode. If deletion were by hand,
> the
> > > > > summary
> > > > > > > would also serve to keep tabs on which groups were in fact
> > > empty
> > > > > and
> > > > > > > needed to be removed. But as I said, I don't think a
> summary
> > > > > feature is
> > > > > > > necessarily something grouper-loader should have.
> > > > > >
> > > > > --
> > > > > Dr. Loris Bennett (Mr.)
> > > > > Freie Universität Berlin
> > > > > Zentraleinrichtung für Datenverarbeitung / Computer Center
> > > > > ZEDAT
> > > > > Compute & Media Service
> > > > > Fabeckstr. 32, Room 221
> > > > > D-14195 Berlin
> > > > > Tel ++49 30 838 51024
> > > > > Fax ++49 30 838 56721
> > > > > Email 
> > > > > 
> > > > > Web www.zedat.fu-berlin.de
> > > > >
> > >
> >