Ref Cursor vs XML
We are writing a web app which will access a Oracle database (11gR2). The web layer will be a mixture of Java/JSF and PHP (different teams tasked with different stuff).
This is a new development. We are aiming to have the app do all SQL operations through store procedure interfaces, i.e. no direct SQL.
One thing under consideration is returning query results - XML or Ref Cursor.
The idea of using XML is appealing because it would allow the web coders to slice and dice the data as they like, and would return all the data in one go. The only problem with it we see is that if a dataset is very large, returning and processing that huge lot of data will affect end-to-end performance.
This is a new development. We are aiming to have the app do all SQL operations through store procedure interfaces, i.e. no direct SQL.
One thing under consideration is returning query results - XML or Ref Cursor.
The idea of using XML is appealing because it would allow the web coders to slice and dice the data as they like, and would return all the data in one go. The only problem with it we see is that if a dataset is very large, returning and processing that huge lot of data will affect end-to-end performance.
0