This section describes required and behavioral changes to the
Endeca Query Language (EQL).
The following changes in EQL occurred in this release:
- Addition of grouping sets,
including
CUBE,
ROLLUP, and their helping functions
(GROUPING,
GROUPING_ID,
GROUP_ID).
- ORDER BY
now can take an arbitrary expression.
- TO_DATETIME
now supports a numeric interface, in addition to the previous string interface.
- The
CONCAT,
STRING_JOIN,
TO_STRING, and
SUBSTR functions were added.
- GROUP BY
ManagedAttr:level is no longer supported. The
ANCESTOR function can be used as an alternative.
- Optional depth to
GROUP BY elements is no longer allowed.
- Duplicate
GROUP BY restriction is removed.
- GROUP and
GROUP BY now respect NULL values and do not discard
records with NULL values.
- Aliasing is not longer
required for attributes.
- SELECT *
from intermediate results is added.
- IS_ANCESTOR
and
IS_DESCENDENT functions (which previously worked only
with managed attributes) now also work with standard attributes.
- Parsing and validation
messages from EQL processing have been localized to provide international
language support.
- Using single quotes with
string values. When using string value comparison operators, make sure that you
use single quotes around the text value field. When using numeric value
comparison operators, do not use quotes of any kind around the value field.
- Escaping special XML
characters. If you are making direct queries against the Conversation Web
Service (for example, by using the soapUI tool), you may need to escape some
XML characters to prevent parsing errors. For example, you should use the
< escape character instead of the
< (less than) character.
Note: This list only announces the changes but does not describe them in
detail. For detailed information on EQL, see the
Oracle Endeca Server EQL Guide.