<number of> physical reads for user task <task_ID>, <number of> commands, application pid <pid>
If a table is accessed for the first time, or if it has not been used for a long time, and was therefore displaced from the data cache, then this situation is not a problem.
However, if this does not explain the read activity, check the hit rate for the data cache, and increase the size of the data cache, if necessary.
Also make sure that the SQL statements specified by the application do not read significantly more data than is necessary for processing, because of poor search strategies, for example.
If the special database parameter LRU_FOR_SCAN has the value yes, then table scans use only 10% of the data cache for caching the table. This could mean that it was not possible to accommodate the entire table in the data cache, and the next scan has to read it physically again.