8 Replies Latest reply on Jul 18, 2016 9:31 AM by David Li

    Regex does not calculate correctly after extract refresh

    jessica.muse.0

      I've been googling and searching the forums for someone having a similar issue, but to no avail. 

       

      Basic info to start: using Tableau 9.0 (and this is controlled by my company, so upgrading can't be the only answer, please) and connecting to an Oracle database (again, don't know the version, etc., as that's company-controlled).  I keep the data source on an extract, and the data within the source is updated nightly.  The report I've generated needs to be updated daily, so the extract is on an automatic refresh schedule.

       

      The issue:

      I'm using a REGEXP_MATCH to find patterns within a field in my database, we'll call it [Text].  The basic overall pattern is ABC_ABC 1232.ABC-ABC_ABCDEFG ABC.  At a more basic level, it's two to seven upper case letters, followed by an underscore, followed by any combination of letters, numbers, and punctuation (other than underscores), followed by an underscore, followed by any combination of letters, numbers, and punctuation (other than underscores).  Now, I have created this regex and it's completely functional.  My reports generate, the data is validated, and life is generally great....until I refresh the extract.  Once the extract is refreshed, the REGEXP_MATCH no longer works.  About half of the [Text] fields are appropriately marked as matching my pattern.  If I edit the REGEXP_MATCH in my calculated field, and just change one character (such as changing {2,7} to {2,6} or a {2,} to a {1,} then it's completely functional again.

       

      Current iteration of the REGEX, just in case (and I have used the \w, \W, \d, {1,} instead of the +, and some of the other shortcuts as well, and the issue is still the same):

      REGEXP_MATCH([Text],"^[A-Z]{2,7}\_([A-Za-z0-9\-\@\\\s]+)\_([A-Za-z0-9\-\@\\\s]+)$")

       

      Any help would be appreciated.  Sorry I cannot provide an example workbook.  As well, I don't think it's a REGEXP or workbook issue, but perhaps a functionality issue.

       

      Thanks in advance.