SQL/MX 2.x Reference Manual (H06.04+)

SQL/MX Utilities
HP NonStop SQL/MX Reference Manual—540440-003
5-35
Considerations for import
°
You can specify that the column be given the default value (for the column) by
starting the input data specification with the character “D” or “d”. import ignores
anything following that and preceding the end of the input field.
°
If the first character of the input data is anything other than “N”, “n”, “D”, or “d”,
import ignores the first character. import uses the data beginning in the next
character position up to the end of the input field as the value for the column.
•
The col=column_name,start,length line in the format file specifies the byte offset
(within each row) and the length (in bytes) where you specify the value for that
column. If you specify Y for the nullDefault_flag for a particular column, the leading
indicator character (where you might put “N”, “n”, “D”, or “d”) is part of the
specification of the value for the column, so the start value should be the byte
offset where that leading indicator character is found, and the length value should
include the leading indicator character, because it is part of the specification for the
value.
Parallel Load for import
Use parallel load when the destination table is partitioned. When using parallel load:
•
Some data types require more CPU time during import and therefore parallel load
would be a benefit.
•
Sorting input data by storage key results in faster import time.
•
More processors improve parallel load performance.
You cannot import into one partition in parallel. You receive a locking error if you have
two instances of import loading the same partition.
You can perform a parallel load in two ways:
•
Run multiple instances of import—one for each partition in the destination table—
to load data into a partitioned table by using a single input file. For each import
command, specify the number of input rows (or records), the number of the first
record to import, and the transaction size. The number of the first record to import
begins with zero.
For example, suppose that you partition the EMPLOYEE table into three partitions.
The first partition begins with 0 for the employee number, the second partition
begins with 3000 for the employee number, and the third partition begins with 5000
for the employee number.
Note. For better performance, you must specify the transaction size. If you do not, and
other import processes are running on the same table, import issues a warning regarding
performance. If the table is empty and does not have any indexes, the first import process
might turn off auditing for all partitions. Turning auditing off enables the first import process
to use a fast loading technique. However, performance of the parallel import processes on
the remaining partitions is affected, because auditing for all partitions is off, and the table
contains data as a result of the first import operation.