Creating the Bank

Initially new banks are created in some temporary file, say SCRATCH.DAT (see 8.1.3), following standard ZEBRA TZ (Titles Handling) format. See the ZEBRA TZ Reference Manual for full details, but the following example will serve to illustrate the most important features:-

*LOG
*---  Standard Titles File: geometry.dat               11 Aug 1995
*---                 Banks: GEDP, GEDS, TGEO
*LOG OFF
*---
*---    Geometry (GEO):  Detector Geometry control bank.
*---
*---    Contact:  N. West  Oxford.
*---
*---
*DO   TGEO    1 -i(31I 2F -I) -n20 #.  Detector Geometry control bank.
#.
#.
#.       Standard Database Header
#.
19750101        0 20380516   223319  #.  1..4   
       0        0        0           #.  5..7   
       0        0        0           #.  8..10  
19750101        0 20380516   223319  #. 11..14  
       0        0                    #. 15..16  
4*0                                  #. 17..20  
10*0                                 #. 21..30  
#.
#.    End of Standard Database Header
#.
#.    User Data.
0                            #. Partial fill flag
20.                          #. Partial fill D2O level.
50.                          #. Partial fill H2O level.
0                            #. Debug flags.
6*0                          #. Spare.

Note that:-

• The *LOG switches on output to the log file and *LOG OFF stops the logging.

• The *--- lines are comment lines used between titles banks. They are written to the log file.

• Each bank starts with a *DO command followed by the bank name and number. This is followed by the compulsory (for SNO!) immediate format -i.... The format specifier '-i(31I 2F -I)' in the example above indicates that the first 31 words are integer, then 2 floating point numbers, and the rest are integer. See the titles banks distributed with SNOMAN for examples of more complicated format specifiers. For full details see the chapter 'Formats for the text input to TZINIT' in CERN's ZEBRA manual or:-

  Companion/ZEBRA: A Beginner's Guide/Bank Format

  Note: Formats are essential for SNO to permit the bank to be stored in the SNO database. It is also essential to limit lines to a maximum of 80 characters.

  The -n followed by a number gives an upper limit on the bank size. The default upper limit is 2000 so isn't required in this case and is only shown as an illustration. Here is an example where the -n isn't optional:-

  *DO   GEDP  462 -i(31I -I) -n48010
  

  ZEBRA needs to know the maximum size before it reads the file; it lifts the bank and then stores data as it reads. There is not need to be precise so long as you don't underestimate; ZEBRA will trim away the excess afterwards. If a bank overflows then ZEBRA will fail with a:-

  "ZFATAM. !!!!! Going to ZFATAL for TZINIT fails.
  

• #. introduces an optional embedded comment. Don't miss out the full stop! Note that they are different to comments between titles banks (which are *---).

• Blank lines are allowed except between the *DO and the first data item.

• Each line can have none, one, or more data items.

• The system decides the type (integer, floating point or hollerith) by the format e.g. 3 is an integer but 3. is floating point. For hollerith use single or double quote stings e.g. 'ABCD' or "A LONG STRING". Hollerith data takes up as many words of a bank as are needed to store it.

  Note: the format is not taken from the supplied format

  It is essential that the format supplied exactly matches the bank contents, as it is used to translate the data in the bank into a machine-independent format for distribution to the various platforms throughout the collaboration. If not, the contents may be damaged when store and retrieved from the database!

• Binary, Octal and Hexadecimal integers may be input by preceeding it with #B (binary), #O (octal) or #X (hexadecimal) e.g.:-

  #B100001 #O017777 #X177AB

• The file can contain any number of banks, a bank is terminated by the EOF or the *DO of the next bank.

Titles files should only be used to input only those banks which have been editted by the user; default titles banks will be read in binary format from the SNO database. However, before new banks (titles) can be entered into the database, they will have to be allocated a directory and alias. Please contact database people (database specialist?) and discuss your requirements with them.