Last update: Edit

1 Introduction

The behavior of Mendix using an Oracle database has some minor differences when compared with using a PostgreSQL database. These differences are documented below.

2 Unlimited and Very Long Strings

The majority of differences between PostgreSQL and Oracle are in how they handle very long, or unlimited length, strings.

2.1 Comparison Functions

Oracle does not support unlimited strings or strings with a specified size greater than 2000 characters when using the equal (=) or not equal (!=) operators in XPath constraints. However, it does support functions including contains(), starts-with(), and ends-with().

2.2 Sorting, Grouping and Aggregating

It is not possible to sort, group, or use aggregate functions such as count() on unlimited strings or strings with a specified length greater than 2000 characters. This is because such long or unlimited strings are implemented with the data type CLOB. Consider decreasing the length of the string attribute or removing it from data grids.

2.3 Selecting DISTINCT Attribute

Selecting DISTINCT attributes of the string type with a size greater than 2000 characters is not supported by Mendix due to a known Oracle limitation of selecting DISTINCT columns with a CLOB data type. If you run into this limitation, you may encounter an exception in the logs with a message like this: Error Msg = ORA-06502: PL/SQL: numeric or value error: character string buffer too small.

3 Required Privileges

The following are the minimum privileges that have to be set for access to your Oracle DB:

  • Create tables

  • Drop tables

  • Create indexes

  • Create sequences

  • Create, read, update, and delete data

These are the requirements for a Mendix Database. The Database Administrator can specify the other access and permissions that are necessary to access the data.