Getting on-server datasource ID from workbook connections #825

Activity

1. thesketh commented on Mar 29, 2021

   

   thesketh

   on Mar 29, 2021

   Last edited by thesketh

   Author

   More actions

   I've been able to do this by parsing the datasource names from the workbook XML, and then matching these with the datasources from the server. Quick explanation:

   
   Tableau workbooks (.twb files) are XML documents, which contain references to the datasources on the server:

   <workbook>
     ...
     <datasources>
       <datasource>
         <repository-location id='datasource-name-here' .../>
       </datasource>
       ...
     </datasources>
     ...
   </workbook>

   If there are two datasources with the same name on-server (e.g. in different projects), these will have a 14 character numerical ID as a suffix:

   datasource-name-here
   datasource-name-here_01759868783642
   

   I parsed these with xml.etree.ElementTree. To link the datasources from the server to these XML names, I used the following function to format the on-server datasource names (TSC.DatasourceItem.name, which I obtained by looping through all the datasources). The regex pattern was obtained by observing the behaviour of our datasource names - I may have missed some different behaviour for other characters:

   def format_source_name(name: str) -> str:
       """Format a datasource name to match the name from the XML."""
       return re.sub(r"[^A-Za-z0-9_\-\.]", "", name).replace(".", "_")

   If these matched with the XML name (except for the ID suffix), I then downloaded the datasource and parsed the XML to ensure that the ID matched:

   <datasource>
     <repository-location id='datasource-name-here_01759868783642' .../>
     ... 
   </datasource>

   
   

   So, in order to identify the datasources connected to a workbook:

   1. Download the workbook's XML (this might be a zip file (.TWBX), so it may be necessary to extract this to obtain the .TWB XML file)
   2. Parse the datasource names (and possibly IDs) from the XML file.
   3. Loop through the on-server datasources and identify matches by formatting the datasource names using format_source_name.
   4. Where the formatted name matches the name from the workbook XML file (ignoring the ID), download the datasource's XML (this might be a zip file (.TDSX), so it may be necessary to extract this to obtain the .TDS XML file)
   5. Ensure that the name from the repository-location tag (including the ID) matches the workbook.
2. thesketh commented on Mar 29, 2021

   

   thesketh

   on Mar 29, 2021

   Author

   More actions

   This is obviously non-optimal, a better solution would be very appreciated!
3. siyingwang-shc commented on Jul 15, 2021

   

   siyingwang-shc

   on Jul 15, 2021

   More actions

   Hi there, I encountered exactly the same issue, any new progress? Thank you!!!
4. jharris126 commented on Sep 6, 2021

   

   jharris126

   on Sep 6, 2021

   Last edited by jharris126

   Contributor

   More actions

   The best way I know to do this with TSC is to use Tableau's Metadata API and throw the GraphQL query string into the metadata endpoint in TSC. Before the Metadata API, I used to parse the XML as well, but this is where the Metadata API shines!

   graph_query = '''{
     workbooks {
       name
       luid
       upstreamDatasources {
         luid
         name
         __typename
       }
     }
   }'''
   
   with server.auth.sign_in(auth):
       wbs = server.metadata.query(graph_query)['data']['workbooks']
   
       for wb in wbs:
           for ds in wb['upstreamDatasources']:
               print(f"{ds['__typename']} datasource: {ds['name']} is referenced by workbook: {wb['name']}.")
5. jacalata

   added

   enhancement

   on Nov 8, 2021
6. jacalata

   added

   Server-Side Enhancement

   on Mar 25, 2023
7. github-actions commented on Aug 6, 2026

   

   github-actionsBot

   on Aug 6, 2026 – with GitHub Actions

   More actions

   This issue has been automatically marked as stale because it has not had recent activity. It will be closed in 14 days if no further activity occurs. If this is still relevant, please comment to keep it open.
8. github-actions

   added

   stale

   on Aug 6, 2026
9. jacalata commented on Aug 6, 2026

   

   jacalata

   on Aug 6, 2026

   Contributor

   More actions

   Closing — this actually works today. ConnectionItem.datasource_id and ConnectionItem.datasource_name are populated when the connection is loaded via workbook.connections after server.workbooks.populate_connections(workbook):

   workbook = server.workbooks.get_by_id(wb_id)
   server.workbooks.populate_connections(workbook)
   for conn in workbook.connections:
       print(conn.datasource_id, conn.datasource_name)
   

   The attribute has been on ConnectionItem since the initial commit; the docstring clarifying its meaning was added in PR #1526.
10. jacalata

    closed this as completedon Aug 6, 2026