Tech and T-Accounts

Modify Registry so twb and sql files show in preview

Use this approach to modify the Registry so that other file types are previewable as though they are text files.

Approach found here:

Open the registry editor (regedit in the start menu search) and navigate to computer\HKEY_CLASSES_ROOT\.FILE_EXT, where FILE_EXT is the extension you want to add, such as .twb  or .tps or .tds or .sql

Make 2 new string values (REG_SZ) under the file extension's key, Content Type andPerceivedType
(note, only the PerceivedType is needed if Content Type already exists -- also note that PerceivedType has no spaces) 

Set the value of Content Type to text/plain and PerceivedType to text

For other Tableau items, like twbx, tdsx that are zip folders, add a PerceivedType as compressed in order for Windows File Explorer to provide insight into those files


Tableau Server 'All Users' missing users

In an interesting twist, the system generated 'All Users' local group didn't have all the users that are listed in the Users tab for a site.  From running this SQL against the Postgres tables/views, confirmed that the users table and the group_users table were not aligned:

select case allusergrpmember when 1 then 'In All Users' else 'Not in All Users' end as GroupMembership, count(*) as ID_Count
select, friendly_name, name, sum(case group_id when 2 then 1 else 0 end) as AllUserGrpMember from _users
 left join group_users gu on gu.user_id =
  where site_id = 1 
 group by, friendly_name, name
 -- having sum(case group_id when 2 then 1 else 0 end) = 0
 order by friendly_name
 ) t
 group by allusergrpmember
So, the intermediate fix was to create a second 'All Other Users' local group to add the missing members.  Tableau was notified of the issue.

Tableau Server - Bad behavior

What constitutes bad behavior on the part of Tableau Server users?  Here's my view, what's yours?

  1. Old reports that aren’t used yet still run extracts.
  2. Failed extracts, especially repeat offenders on PROD.
  3. Huge number of subscribers to reports which take minutes to render.
  4. DEV versions on PROD server that aren’t actively being UAT.
  5. Repetitive extracts embedded in workbooks rather than as a datasource.
  6. Huge extracts when someone has access to a more cost effective db solution.
  7. Custom SQL when a Tableau build with referential integrity could have been done.
  8. Thirty or so connections in a single workbook.
  9. Using Projects instead of Sites.
  10. Governing AD groups by an organization distant from the managers of the AD group.
  11. Leaving departed users licensed on a server.
  12. Poor Tableau performance from attempting to design according to another product’s ‘best practices’.
Page 3 of 7 << < 1 2 3 4 5 6 7 > >>
Copyright© Brad Earle
Site Map | Printable View | © 2008 - 2021 KB Earle Associates LLC | |