Use the information in this section to help troubleshoot caching for a Joiner transformation.
The following warning appears when I use the cache calculator to calculate the cache size for a Joiner transformation with sorted input:
CMN_2020 Warning: If the master and detail pipelines of a sorted Joiner transformation are from the same source, the Integration Service cannot determine how fast it will process the rows in each pipeline. As a result, the cache size estimate may be inaccurate.
The master and detail pipelines process rows concurrently. If you join data from the same source, the pipelines may process the rows at different rates. If one pipeline processes its rows faster than the other, the Integration Service caches all rows that have already been processed and keeps them cached until the other pipeline finishes processing its rows. The amount of rows cached depends on the difference in processing rates between the two pipelines.
The cache size must be large enough to store all cached rows to achieve optimal session performance. If the cache size is not large enough, increase it.
This message applies if you join data from the same source even though it also appears when you join data from different sources.
The following warning appears when I use the cache calculator to calculate the cache size for a Joiner transformation with sorted input:
CMN_2021 Warning: Increase the data cache size if the sorted Joiner transformation processes master rows that share the same key. To determine the new cache size, divide the existing cache size by 2.5 and multiply the result by the average number of master rows per key.
When you calculate the cache size for the Joiner transformation with sorted input, the cache calculator bases the estimated cache requirements on an average of 2.5 master rows for each unique key. If the average number of master rows for each unique key is greater than 2.5, increase the cache size accordingly. For example, if the average number of master rows for each unique key is 5 (double the size of 2.5), then double the cache size calculated by the cache calculator.
