STAR   Computing Tutorials main page
STAR Databases: Navigation

Offline computing tutorial Maintained by Jeff Porter
Last modified Fri Jul 03 00:19:00 1998

Navigation

Consistent use of all STAR database data requires that the various data have a well understood navigation scheme. This scheme relies on references by run-number, event-number, time-stamp, & version-names. The general picture of how these data are tied together is shown below. Not all databases are depicted (e.g. Geometry) but the basic access keys between database types are indicated.

See also the descriptions below for,


TimeStamp

The TimeStamp that is stored both on the DAQ Raw data files and in the Database is the standard unix timestamp (unsigned int). The value is relative to the date, 1969-12-31 19:00:00 (unix-time=0), and has a unique range up to the date, 2037-12-31 23:59:59 (unix-time=2145934799). MySQL has a storage type called Date-Time for which the storage is unix-time while the display is done in the ascii string format noted here,

"YEAR-MO-DY HR:MN:SE"
Both formats of the timestamp are supported in the C++ API such that if one format is used to set the request or store times, the other format form is filled immediately.


Version Name

The Version Name is a 128 char string (127+'\0') that identifies a particular version of data instances apart from any default version. In short, it is one of the primary keys associated with the lookup of database data. For example, two calls (below) prepare data for store (or request) the C++ API,

StDbTable* mytable1 = myNode->addDbTable("aTableName");

StDbTable* mytable2 = myNode->addDbTable("aTableName","aVersionName");
differ only in that subsequent "store" ("fetch") calls will write (retrieve) different versions of the table "aTableName".


Run & Event Number


Servers & DataBases