EBUS_JDBC_MULTIWRITE
|
Meaning: |
With
Index Files of type JDBC WRITE instructions take considerable computing time,
because each WRITE instruction requires a separate SQL instruction to be
executed.
If
the variable EBUS_JDBC_MULTIWRITE is set to 1, those
WRITE instructions which create new data records are not executed, but
the data records are cached in memory buffer program-internally until the
function JDBC_MULTIWRITE(<Index File Name>) is executed. The
function JDBC_MULTIWRITE writes all cached data records with one SQL
instruction into the database. This increases performance significantly. EBUS_JDBC_MULTIWRITE
does not have any effect with existing data records that are overwritten with
WRITE. The
advantage is the speed increase, the disadvantage is that errors might occur
only when the function JDBC_MULTIWRITE is called (e.g.
if a data record that did not exist before has meanwhile been created by
another program). In the case of an error, since several data records are
rewritten with one single SQL instruction, the data records after the faulty
data record are not saved anymore and are therefore lost.
EBUS_JDBC_MULTIWRITE
can, e.g., be used to accelerate the program execution time if many data records
are to be imported into a SQL database. Note: You should make sure that
no more than 100 data records per Index File are cached in the memory buffer.
With more than 100 data records the gain in performance decreases again since
the administration effort increases.
The
function JDBC_WRITE_MODE is a further means of increasing
performance.
This
variable is set to 0 at the start of the program. It only has an effect with
JDBC Index Files. |
|
Changeable: |
Yes |
|
See also: |