Database Reference
In-Depth Information
Finding the source bottleneck
Performance can also be impacted on the source side. Various factors can hamper the per-
formance on the source side, just like we discussed for the target side.
Using thread statistics
As discussed in the
Finding the target bottleneck
section, use thread statistics to find the
source bottleneck. Refer to this section to use thread statistics.
Test mapping
Using a pass-through mapping can help you identify whether the bottleneck is on the
source side. Configure a test mapping to have only Source, Source Qualifier, and Target.
Consider that you have a mapping with one source, one target, and 10 transformations, and
the time taken to process the data through the mapping is 60 seconds. This means combin-
ing the source, target, and transformations is taking 60 seconds.
Create another mapping by eliminating all the transformations and run the process and
check the time. Suppose the time taken to complete the process is now 50 seconds. As
compared to the combined time of the target and transformation, which is 10 seconds, 50
seconds is relatively high, which can indicate that we have a source bottleneck.
Using a Filter transformation
You can use a Filter transformation to check whether the source has a bottleneck. Put a Fil-
ter transformation in front of the Source Qualifier and set the condition to
False
, that is,
don't allow any record to pass through the filter in the mapping. Using this method, you can
compare the runtime of the process with and without the filter condition. This way, you can
identify whether the source has a bottleneck.
Checking the database query
This is another simple method to find the source bottleneck if you are extracting the data
from the table. When you use the Source Qualifier to read the data from the database table,
Integration Service writes a query to extract the data. Copy the query and run the same
query at the database level in the SQL developer or a similar tool. Compare the time taken
by the query to get executed at both places, which can give you an idea whether the source
has a bottleneck.
