Nintex Connect

The Official Community Site for Nintex Software Products, including Nintex Workflow 2007, Nintex Podcast 2007 and Nintex SmartLibrary.
Welcome to Nintex Connect Sign in | Join | Help
in Search

Execute SQL for Oracle returns invalid characters

Last post 02-09-2010, 2:45 AM by StevenB. 3 replies.
Sort Posts: Previous Next
  •  02-07-2010, 1:49 AM 8736

    Execute SQL for Oracle returns invalid characters

    truncate table TABLE1;
    truncate table TABLE2;

    This SQL code returns Error performing database operation. ORA-00911: invalid character.

    By trial and error I've determined that the ; is causing the error.  Does this mean that if I need to truncate 10 tables then insert data into those same 10 tables I need 20 Execute SQL functions in my workflow?  Is this correct?

  •  02-08-2010, 12:52 PM 8751 in reply to 8736

    Re: Execute SQL for Oracle returns invalid characters

    Hi Steve,

    I am not familar with Oracle syntax, but is the semi colon how you would usually separate statements in oracle? The Execute SQL action doesn't do any processing of the statements (except replacing the context data), it just treats the whole text as one string and passes it to the database provider. It is the database provider that is reporting an error.

    One possibility though, are your two commands 'hard coded' or are you using reference data to construct them?

    regards

    Jaems

  •  02-09-2010, 2:19 AM 8763 in reply to 8751

    Re: Execute SQL for Oracle returns invalid characters

    Hi,

    Thanks for the reply.  Yes, the semicolon is how you separate commands in Oracle SQL.  No, the commands are "hard coded," no references from the workflow.  Although the one command that does use workflow references works fine as it's only 1 SQL command in an Execute SQL box with no semicolon.  I've also noticed I cannot use the command "host" (which would allow me to call an executable like imp.exe or exp.exe allowing me to run imports or exports) or a "@" (which is used to start a line that calls a .sql file from the file system to process).  All-in-all, the Execute SQL does seem to process, or at least, parse and validate the SQL code, but I agree the error I posted does come from the database.

     The process I'm trying to build is for a refresh process from our development database to our production database.  I can get by without the "host" or "@" commands, it just makes my workflow longer, but if I can't put more than one command in the Execute SQL function then the workflow becomes so large as to be unmanageable/unusable.

  •  02-09-2010, 2:45 AM 8765 in reply to 8751

    Re: Execute SQL for Oracle returns invalid characters

    James,

    I've confirmed that a command like "truncate TABLE1;truncate TABLE2;" returned the invalid character error.

    This suggests to me that the Execute SQL function is stripping out the CR/LF.

View as RSS news feed in XML
Powered by Community Server, by Telligent Systems