I'm running on server version 8.1.00 201301251157.  Of course, I guess they
may've fixed this in a subsequent patch...

-charlie


On Fri, Mar 28, 2014 at 1:58 PM, LJ LongWing <[email protected]> wrote:

> **
> Charlie,
> All I can say is 'WOW' :).....that's just insane.  Can you provide your
> exact version/patch?
>
>
> On Fri, Mar 28, 2014 at 2:27 PM, Charlie Lotridge <[email protected]>wrote:
>
>> **
>> If you use ARS views, this could be a problem for you.
>>
>> It's probably no surprise that if you modify the SQL table (or view) used
>> as the "table" for an ARS view, and (effectively) rename one of the columns
>> used by one of the ARS view fields, the ARS view breaks.
>>
>> What I discovered, though, is that if you simply go into the view
>> definition modify the "Column" property in the View Information section of
>> the properties for that field to have the corrected column name, it DOES
>> accept this change back in the server, but the server DOES NOT update the
>> data dictionary with this change.  So, a subsequent bounce of ARS will
>> restore the old, incorrect column name.
>>
>> The work-around is to delete the field from the view (cut it), save the
>> view, then paste it back in and correct the column name.
>>
>> This SQL query will reveal any such problems in your system (written for
>> SQL Server but probably pretty easily adapted to other DBs):
>>
>> SELECT s.name, f.fieldName, sv.tableName, vm.extField AS "Bad Column
>> Name"
>> FROM arschema s
>> JOIN field f
>>  ON f.schemaId = s.schemaId
>> AND f.fOption != 4
>> JOIN schema_view sv
>>  ON sv.schemaId = s.schemaId
>> JOIN view_mapping vm
>> ON vm.schemaId = s.schemaId
>>  AND vm.fieldId = f.fieldId
>> LEFT OUTER JOIN sys.all_objects t
>> ON t.name = SUBSTRING(sv.tableName, 1, LEN(CAST(sv.tableName AS
>> NVARCHAR(128))) - 1)
>>  AND t.type in ('U', 'V') -- User Table or View
>> LEFT OUTER JOIN sys.columns c
>> ON t.object_id = c.object_id
>>  AND c.name = vm.extField
>> WHERE c.object_id IS NULL
>>
>> Another interesting and impactful bug related to this (and, in fact, how
>> I originally discovered there was a problem with my view form), is that if
>> you export the bad view form definition and import it into another server,
>> it WILL import with only a warning:
>>
>> Warning (69): Creation of one of the SQL views for the form failed within
>> the SQL database.
>> Form is created successfully, but the SQL view is not in place.: <view
>> name>
>>
>> where <view name> here is the name of the courtesy SQL view that Remedy
>> creates for each form (e.g. the "Customer" SQL view ARS will create for
>> your Customer form).
>>
>> For view forms, the "T table" itself is a SQL view defined on the
>> underlying SQL table (or view), and the courtesy SQL view that ARS then
>> creates is simply another SQL view on top of the "T table" view.  So it
>> would appear from the warning that only this final SQL view creation is
>> failing.  However, a SQL log I captured during such an import revealed that
>> the creation of the "T table" view itself was also failing.  But apparently
>> this is not being noticed by ARS, because it does go on to create the ARS
>> data dictionary stuff for the form (i.e. the form WILL appear in your list
>> of forms on that server).  However, any attempt to actually use the form
>> fails with a cryptic message such as:
>>
>> The SQL database operation failed. :
>> Invalid object name 'T435'. (SQL Server 208) (ARERR 552)
>>
>> Anyway, sorry for the complex description here and hopefully I haven't
>> left anyone interested too confused...let me know if you have any questions.
>>
>> I don't currently have a client, so I have no active support id through
>> which I can submit a ticket on this.  So, anyone, feel free to do so if
>> this affects you.
>>
>> Thanks,
>> Charlie
>> _ARSlist: "Where the Answers Are" and have been for 20 years_
>
>
> _ARSlist: "Where the Answers Are" and have been for 20 years_

_______________________________________________________________________________
UNSUBSCRIBE or access ARSlist Archives at www.arslist.org
"Where the Answers Are, and have been for 20 years"

Reply via email to