SUPPORT / SAMPLES & SAS NOTES
 

Support

Usage Note 37979: Failed SAS® OpRisk VaR dataload

DetailsAboutRate It

The information in this SAS Note augments the list of data tables, variables, and formats that is available in:

  • SAS OpRisk VaR 3.2 User's Guide, Chapter 3, "Loading Data into SAS OpRisk VaR"
  • SAS OpRisk VaR 4.1 Help, "Appendix B: Structure of Source Data in SAS OpRisk VaR"
  • SAS OpRisk VaR 4.2 Administrator's Guide, "Appendix 2: Structure of SAS OpRisk VaR Source Data"
  • SAS OpRisk VaR 5.1 Administrator's Guide, "Appendix 2: Structure of SAS OpRisk VaR Source Data"
  • SAS OpRisk VaR 6.1 Administrator's Guide, "Appendix 2: Structure of SAS OpRisk VaR Source Data"

For successful dataloads into SAS OpRisk VaR, use the following guidelines and information.

  • Ensure that the minimum, internal data tables are included in the data load:
    • Standard Business Line Structure
    • Standard Risk Category Structure
    • Internal Business Line Structure
    • Internal Risk Category Structure
    • Loss Event Detail
    • Starting with the 6.1 release, the Global Range data set is also required.
  • Ensure that all required variables are included in these tables.


  • Ensure that the rows containing the Name, ID, and Path Component variables for the cross-classification dimensions are in the proper order. Rules are as follows.
    • The ID must be an integer and unique in the data set because it is used as a primary key.
    • The Name row (top level) must be listed before the rows containing the Path_Component names (lower levels). E.g., Name=A, where "name" refers to a specific level of a cross-classification dimension (standard business line, internal business line, standard risk category, internal risk category, process, legal entity, etc). See row 1 in the image below.
    • Path_Component_1 must be first listed as a Name. E.g., Name=C, Path_Component_1=A. See row 3 in the image below.
    • Anything that is in Path_Component_2, must be first listed as a Name in a record with the same Path_Component_1. E.g., Name=G, Path_Component_1=A, Path_Component_2=C. See row 7 in the image below.
    • Anything that is in Path_Component_3, must be first listed as a Name in a record with the same Path_Component_1 and Path_Component_2. E.g., Name=J, Path_Component_1=A, Path_Component_2=C, Path_Component_3=G. See row 10 in the image below.
    • And so on for any further path components. The resulting dimension hierarchy could be as shown in the following image.
    • Input Data Rows for an Example Hierarchy
    • Failure to follow these rules can cause a dataload failure, and a message like the following occurs:
      STATUS Obs 0000013 - The path/id could not be located for (_path_component).
  • Ensure that if values are required for the required variables that these values
    • are included.
    • have the proper formats. As examples:
      • If numeric values are required, then character values should not be used.
      • The date column is a numeric field (that is, number of days from January 1, 1960), and formats can be applied, for example mmddyy10. However, if character strings are used, the dataload fails.

    • are appropriate. As examples:
      • There are 2 calculations for the Loss Event Detail, which must hold for the dataload to be successful.
      • The date variables must conform to the required form; for example, Last_Date_of_Event Date should be greater than or equal to the dates listed for First_Date_of_Event.
      • If loading date values from Microsoft® Excel spreadsheets, the numeric dates values might be inappropriate. For example, when Excel shows the numeric date as 39022 for 11/1/06, in SAS the number is 39022 - 21916 = 17106. Subtract 21916 from the Excel numeric dates to get the SAS numeric dates. This difference occurs because Excel numeric dates start at JAN 0, 1900 as 1, and SAS numeric dates start at JAN 1, 1960 as 1, which is a difference of 21916 days.

  • If values are not required for the required variables, then:
    • these data values can be missing.
    • however, if specific values are provided, then the tables to which they refer must also be loaded, and these tables must

      • have the proper formats. For example, if numeric values are required, then character values should not be used.
      • have the required variables.
      • have the specific values that are referred to by other data tables.

For example, the Loss Event Detail data table is required for a minimum, internal data load. The variable Country_of_Incident_ID is required. However, values are not required. If values are not provided, then the dataload can be successful. If values are provided, then the Country data table must also be loaded, and there must be matching (integer) values for the Country_ID variable -- there must also be values for the (character, 50) Country_Name variable -- otherwise, the dataload fails. The Country data table provides a mapping between the Country_ID and the Country_Name variables.

  • If the dataload is an incremental load and if _all_ of the data tables that have reference variables have been loaded previously, these data tables do not have to be re-loaded for the incremental load. However, if new specific values for reference variables are loaded, then the reference data tables must also be included in the incremental dataload.


  • If Risk Factor (KRI) data are to be loaded, at least 3 observations are required. Otherwise, the Risk Factors cannot be included in the distribution fitting. Note that Risk Factor data can be entered at any valid interval. Often monthly observations are used. If Risk Factor data are quarterly, then you should enter one value per quarter per business line, and the data are automatically expanded to monthly and then forecast out one year past the end date of the losses for them.


  • Data from SAS® Enterprise GRC (and the earlier SAS® OpRisk Monitor) export data tables might require modification before they can be loaded.
    • Use a SAS DATA Step or other SAS Software procedures to prepare these data tables. For example, you can use a SAS DATA Step to include SAS OpRisk Monitor integer data in the CLASS_1 - CLASS_10 columns of the Loss Event Detail table.
    • The prepared data sets must conform to the required structure and format for the dataload into SAS OpRisk VaR to be successful.

  • Ensure that data from SAS Enterprise GRC (and the earlier SAS OpRisk Monitor) export data tables are supported in the SAS OpRisk VaR data tables.
    • In general, SAS Enterprise GRC supports more dimensions than SAS OpRisk VaR, and the SAS OpRisk VaR dataloaders ignore the data table columns that they cannot load.
    • Some SAS Enterprise GRC export data tables might have similar names to data tables that are supported by SAS OpRisk VaR, but still cannot be loaded. For example, the SAS Enterprise GRC scenario tables are different in structure, format, and usage. To use scenarios in SAS OpRisk VaR, the scenario tables must conform to the supported structure and format as specified in the SAS OpRisk VaR documentation listed above.

Notes:

  • Each SAS OpRisk VaR dataload is saved as a version, only one data version per SAS OpRisk VaR project.
  • SAS Note 19893 "ERROR: Write access to member SASUSER.CANCEL_LOAD.DATA is denied"
  • SAS Note 20759 Loading user-defined covariance matrices from a server
  • SAS Note 31812 SAS log output window is empty when calculations are run
  • SAS Note 32524 Documentation incorrectly describes details for rare-event scenario processing in SAS® OpRisk VaR
  • SAS Note 42856 Dataload failure and error if loss_event_detail table is empty
  • Click this link to access SAS OpRisk VaR 3.2.1 Hot Fix
  • Second Maintenance Release Planning for SAS(R) 9.2 SAS OpRisk VaR 4.1
  • Click this link to access SAS OpRisk VaR 4.1_M2 Hot Fix
  • Click this link to access SAS OpRisk VaR 4.1_M3 Hot Fix
  • Click this link to access SAS OpRisk VaR Server 4.2 Hot Fix
  • Click this link to access SAS OpRisk VaR Server 4.21 Hot Fix
  • Click this link to access SAS OpRisk VaR Server 5.1 Hot Fix

Additional Information

For SAS OpRisk VaR dataloads that include export files from SAS OpRisk Monitor, the following SAS Notes might also be of interest.

  • SAS Note 33297 Amounts for near misses are 0 for data loaded into SAS® OpRisk VaR
  • SAS Note 33298 Some impact details are not loaded into SAS® OpRisk VaR
  • SAS Note 33300 Costs are marked as estimates instead of final cost when exporting files from SAS® OpRisk Monitor to SAS® OpRisk VaR
  • SAS Note 33301 Business line and risk category mapping data is missing when exporting files from SAS® OpRisk Monitor to SAS® OpRisk VaR
  • SAS Note 33303 KRI's having multiple observations are loaded incorrectly when exporting data from SAS® OpRisk Monitor into SAS® OpRisk VaR
  • SAS Note 33304 KRI date is not correct when exporting files from SAS® OpRisk Monitor to SAS® OpRisk VaR

Information about the SAS Detail Data Store (DDS) for Banking

  • SAS® Detail Data Store for Banking 2.5 Implementation and Administration Guide
  • SAS® Detail Data Store for Banking 3.1 Implementation and Administration Guide
  • SAS® Detail Data Store for Banking 3.2 Implementation and Administration Guide
  • SAS® Detail Data Store for Banking 3.21 Implementation and Administration Guide
  • SAS® Detail Data Store for Banking 4.6 Implementation and Administration Guide
  • SAS® Detail Data Store for Banking 4.8 Implementation and Administration Guide



Operating System and Release Information

Product FamilyProductSystemSAS Release
ReportedFixed*
SAS SystemSAS OpRisk VaRMicrosoft® Windows® for x64
Microsoft Windows 7
Microsoft Windows 95/98
Microsoft Windows 2000 Advanced Server
Microsoft Windows 2000 Datacenter Server
Microsoft Windows 2000 Server
Microsoft Windows 2000 Professional
Microsoft Windows NT Workstation
Microsoft Windows Server 2003 Datacenter Edition
Microsoft Windows Server 2003 Enterprise Edition
Microsoft Windows Server 2003 Standard Edition
Microsoft Windows Server 2008
Microsoft Windows XP Professional
Windows Millennium Edition (Me)
Windows Vista
64-bit Enabled AIX
64-bit Enabled HP-UX
64-bit Enabled Solaris
HP-UX IPF
Linux
Linux for x64
Solaris for x64
* For software releases that are not yet generally available, the Fixed Release is the software release in which the problem is planned to be fixed.