Database Reference
In-Depth Information
Eliminating the mapping bottleneck
There are various ways in which you can optimize the transformations and mapping.
Using a single pass mapping
Consider a scenario where you have multiple targets to load from the same source; in such
a scenario, avoid creating multiple mappings. You can save significant time by loading
multiple targets in the same mapping. Use a single source and pass the data to different
pipelines and then to multiple targets. This way, you can save the time of reading the same
data multiple times in multiple mappings.
Avoiding data type conversions
Avoid changing the data types across the transformations in the mapping. When you
change the data type, Integration Service takes time to perform the processing. It is always
recommended that you do not change the data type wherever not required.
Unchecking unnecessary ports
Passing unnecessary data through the mapping will hamper performance. It is recommen-
ded that if you do not need certain data, disable the output port of the column so that you
don't pass the data.
Processing the numeric data
Integration Service processes numeric data faster as compared to other data. Try to process
as much numeric data as possible.
Using operators instead of functions
Integration Service processes operators faster as compared to functions. For example, con-
sider using
||
(pipe) in place of the
CONCAT
function to concatenate the data.
Using decode in place of multiple IIF functions
If your logic contains multiple IIF functions, try replacing them using
DECODE
. The
DECODE
function is faster as compared to multiple IIF functions.
