Warehouse Release 3.01.2340

04/04/2006

 

The features in the 3.01.2340 release are:

 

. A new built-in function REPLACE has been added.  The syntax is:

  a source string and replaces characters

 

 

    string = REPLACE(source-string, search-string, replacement-string)

 

 

    REPLACE repeatedly inserts fill-string to the left of source-string

    until the string is of length final-length.  If final-length is

    less than the length of source-string, characters are stripped

    from the RIGHT of source-string to make it length final-length.

 

 

    Replaces all occurrences of search-string in source-string with

    replacement-string.

 

 

    The source-string can be of any character string or wide character

    string type.  If source-string is null, null is returned.  If

    search-string is null, the source-string is returned.  If

    replacement-string is null, all occurrences of search-string are

    stripped from source-string.

 

 

    Since REPLACE is also a standard SQL function, if REPLACE is used

    in the FOR condition of a READ statement on an SQL database

    (Oracle, SQL Server, DB2), the REPLACE is passed on to the database

    engine with the database engine doing the evaluation of REPLACE.

    This may cause unexpected results since the databases may handle

    the corner conditions differently.  For example:

 

 

        On Oracle:     REPLACE('Taurus', 'us', null) results in 'Taur'

        On SQL Server: REPLACE('Taurus', 'us', null) results in null

 

 

    Examples:

        Express                             Result

        REPLACE('BANANA', 'NA', 'S')        'BASS'

        REPLACE('BANANA', 'A', 'N')         'BNNNNN'

        REPLACE('***', '*', '**')           '******'

        REPLACE('I was here', ' ', '')      'Iwashere'

        REPLACE('None', 'X', 'Y')           'None'

        REPLACE('AbcAbc', 'Abcd', 'XYZ')    'AbcAbc'

 

 

. Fixed a problem with the ISNUMP() and NUMP() functions.  In

  previous releases calling ISNUMP or NUMP with an empty

  string would cause Warehouse to abort.

 

 

. On MPE/iX distribution only: A new version of SCNV (version 2.2.0)

  has been included.  Version 2.2.0 of SCNV has the -COG2CSV option

  which converts a Cognos output file to a csv (comma separated values)

  file that is suitable for importing into DataBridger Studio.  (See

  the SCNV documentation file SCNVDOC included with the distribution.)

  Once the csv file has been imported into a DataBridger Studio Format,

  a file may then be opened within Studio using the imported format.

 

. The Warehouse-F build for 64-bit HP-UX (PA-RISC 2.0) will no longer

  be available for download. Please install the Warehouse-H

  32-bit HP-UX (PA-RISC 1.0) instead.

 

 

Release 3.01.2340 is available now on all supported platforms.