There are several types of TFTR data--raw data, waveform, results files, analysis results, and databases. Raw data are written in keyword files . Results from diagnostic programs are usually stored in Ufiles or waveform files; some are stored in keyword files. SNAP, and TRANSP analysis results are in their own formats. The databases are usually stored with INGRES, but some are in an ISAM format. (See LOCUS section 4.3.)
The raw data are archived to tape. The archiving is done automatically and you run GETSHOTS which is described in section 3.1.1 to restore files.
The other types of data are written to magnetic disks and archived on
the optical disk system . These files
are dynamically restored to the magnetic disks as needed. For
example, if you specify a results file that is on the optical disk to
a program like CUPLOT, the program pauses while the file is
retrieved. If you are running interactively, a message is output on
your terminal. When the file has been restored, program execution
resumes. If this wait is excessive you can abort the application with
^Y and the request will still be made. (The magnetic disks
used in the optical disk system are called the read cache.)
To access raw data, waveforms, and results files you need to know the diagnostic identifier . Identifiers are 2 characters long and are called digraphs. This document uses di to denote an arbitrary digraph, as in [TFTR.di]. The list of digraphs is available in the file TFTR_INF:DIGRAF.INF. The present list is given in Appendix B.