Complete event missing in extract

Using 3.8 we are missing complete events in an extracted dataset. The problem occurs in Excel and SPSS output.

Everything else is fine, just this event is missing. No multiple versions of CRF's or whatsoever. Anyone else experienced this problem?

Comments

  • ebsebs Posts: 74
    It may only be missing from the header section. See the following for my and other users who saw the same/similar behaviour as you...

    https://forums.openclinica.com/discussion/15799/crf-missing-from-excel-tab-html-export-header-section

    Cheers
    Eric
  • raoulbraoulb Posts: 4
    Unfortunately this is not the case here. The headers and the data is missing. If I create a dataset with just items from this event, it remains empty, only a list of subjects is exported.
  • ebsebs Posts: 74
    Try creating the extract at study level? If you were at site level and the event wasn't available at that particular site you wouldn't get data.
  • raoulbraoulb Posts: 4
    edited October 2016
    There are no sites, so there is only the study level. Here is what I can see in the ODM export:

    The data for the missing event is present in the <ClinicalData> section. However, the definitions for the particular event, including forms and items are missing from the metadata in the <MetaDataVersion> section.

    This is why the data for this event is not exported to the SPSS and Excel output. Still no clue what is causing this.
  • lindsay.stevenslindsay.stevens Posts: 388 ✭✭
    via Email
    I cant see your screenshot but have you got unicode characters anywhere in
    your study? Or maybe long or duplicate OIDs?
  • raoulbraoulb Posts: 4
    Ok, analysis on the database brought me one step further:

    When crf's are being reused from another study, they sometimes are stored with the study_id from the original study. So the study_id that is stored with the event definition (study_event_definition) differs from the one that is stored in the crf link to the event (event_definition_crf).

    I can see this happening in two of our studies, both single center (no sites defined), and in both cases the extraction does not generate definitions for the events / crf's affected.

    This seems like a bug to me, that is not related to the extraction but to the definition of events.
Sign In or Register to comment.