I have a simple camel route. It polls a directory, reads a text file and splits the content of file line by line. Then it transforms each line (now having lesser no. of characters than original line) and aggregates the transformed lines (messages) into a single message which is written in a CSV file. I am using JdbcAggregationRepository (camel-sql-3.0.0) and database is Oracle 12C.
The problem is that if the file size is 1.67 KB and contains only 3 lines (all lines have more or less same no. of characters), Oracle 12C will use 2 MB TABLESPACE to store the aggregated EXCHANGE (LOB columns) in tables: "repository" and "repository_completed" combined together.
This demand on TABLESPACE has further been observed in case of a file (5.44 MB) containing 10,000 lines which requires nearly 10 GB TABLESPACE in Oracle 12C for JdbcAggregationRepository to work properly.
Is this a totally Oracle 12C LOB handling issue or JdbcAggregationRepository provided out-of-the-box by Camel (3.0.0) has some bug? Has this problem of huge TABLESPACE requirements been faced by anyone - means is it a known issue?
