Previous section   Next section

Namespace-wide Default Collation

Each namespace has a current string collation setting. This string collation is defined for the data type in %Library.String. The default is SQLUPPER. This default can be changed.
You can define the collation default on a per-namespace basis. By default, namespaces have no assigned collation, which means they use SQLUPPER collation. You can assign a different default collation to a namespace. This namespace default collation applies to all processes, and persists across InterSystems IRIS restarts until explicitly reset.
   SET stat=$$GetEnvironment^%apiOBJ("collation","%Library.String",.collval)
   WRITE "initial collation for ",$NAMESPACE,!
     ZWRITE collval
  DO SetEnvironment^%apiOBJ("collation","%Library.String","SQLstring")
  SET stat=$$GetEnvironment^%apiOBJ("collation","%Library.String",.collnew)
    WRITE "user-assigned collation for ",$NAMESPACE,!
     ZWRITE collnew
  DO SetEnvironment^%apiOBJ("collation","%Library.String",.collval)
  SET stat=$$GetEnvironment^%apiOBJ("collation","%Library.String",.collreset)
     WRITE "restored collation default for ",$NAMESPACE,!
     ZWRITE collreset
Note that if you have never set the namespace collation default, $$GetEnvironment returns an undefined collation variable, such as .collval in this example. This undefined collation defaults to SQLUPPER.
If your data contains German text, uppercase collation may not be a desirable default. This is because the German eszett character ($CHAR(223)) has only a lowercase form. The uppercase equivalent is the two letters “SS”. SQL collations that convert to uppercase do not convert eszett, which remains unchanged as a single lowercase letter.
Previous section   Next section