This article explains more about the database size and what the values in EXA_ALL_OBJECT_SIZES mean.
The database size can be calculated on the basis of the following system dictionaries: EXA_ALL_OBJECT_SIZES or EXA_DBA_OBJECT_SIZES.
The dictionary lists all tables and schemas. Please note, the size of an object of the type 'SCHEMA' will be calculated as a sum of the sizes of all the objects in this schema. For views, functions, etc., the size represents the corresponding text size.
RAW_OBJECT_SIZEspecifies the logical object size based on both data types and content. This size is comparable with the size of a CSV-file containing the same data. The value is calculated as a sum of sizes of stored data:
NULL –› 1 byte
fixed size type: depending on the type –› 1-80000 bytes
variable size type (varchar) –› number of bytes the value really uses.
MEM_OBJECT_SIZEspecifies the real size of the database object. The value is calculated as a sum of the following:
a sum of all stored values after compression
structural overhead, e.g. length information for a VARCHAR value
overhead for replication Replication: Table content will be held in RAM on each node for better performance. This applies only to small tables (< 100.000 rows).
Please note, that for a new table some data blocks will be reserved. Therefore, MEM_OBJECT_SIZE of empty or very small tables can be bigger than RAW_OBJECT_SIZE. This does not imply a bad data compression ratio.
These system dictionaries provide you the total size of database objects in the cluster.