Cluster Databases:

  Structure of Cluster Databases:

       [1. MANDT char 3 - client;]
	2. RELID char 2 - area ID;
	3. xxxxx char xx - key (may be padded by system);
	4. SRTF2 int4 4 - line counter of the current cluster (=0...2**31-1);
       [5. user data, e.g. ... PGMID, User ID etc. (several fields);]
	6. CLUSTR int2 2 - length of data in CLUSTD;
	7. CLUSTD varc .. - data cluster (usually length=1000...2000).

  Operations:

	EXPORT <f1> [FROM <g1>] <f2> [FROM <g2>] ... 
		TO DATABASE <dbtab>(<area>) [CLIENT <client>] ID <key>.

	IMPORT <f1> [TO <g1>] <f2> [TO <g2>] ... 
		FROM DATABASE <dbtab>(<area>) [CLIENT <client>] 
		ID <key>|MAJOR-ID <maid> [MINOR-ID <miid>].

	DELETE FROM DATABASE <dbtab>(<area>) [CLIENT <client>] ID <key>.

	IMPORT DIRECTORY INTO <dirtab> 
		FROM DATABASE <dbtab>(<area>) [CLIENT <client>] ID <key>.
		
		DATA DIRTAB LIKE CDIR OCCURS 10 WITH HEADER LINE.

		Field	Type	Description
		--------------------------------------------------------------	
		NAME	CHAR	Name of stored object in cluster.
		OTYPE	CHAR	Type of object:
					F means elementary field
					R means field string
					T means internal table
		FTYPE	CHAR	Data type of object. Structured data types are type C.
		TFILL	INT4	Number of filled lines (for internal tables).
		FLENG	INT2	Length of field or structure.