Presto Administration : Presto Server Administration : Upgrade to New Versions of Presto and Migrate Artifacts : Migrate Mashups That Use RAQL for 3.7
Migrate Mashups That Use RAQL for 3.7
 
Handle In-Memory Store and Load Changes
Handle Column or Function Name Incompatibilities
Updates for Non-Grouping Columns in Queries with Aggregation Functions
Updates for Alias References in Select Clauses
Handle Errors from Derived Schemas
In release 3.7, several updates were made to the Real-Time Analytics Query Language to provide closer compatibility with the SQL 2003 standard. This includes both changes to syntax and changes to some built-in functions. These changes are not backwards compatible.
To migrate to the 3.7 release, you may need to manually update mashups with RAQL statements in the following cases:
Mashups with
May Need Manual Updates
<storeto> or <loadfrom>
RAQL now stores each dataset in its own In-Memory Store rather than storing many datasets in one In-Memory Store. This has significantly changed the syntax and semantics of both the <storeto> and <loadfrom> statements.
See Handle In-Memory Store and Load Changes for migration techniques.
Column or function names that match reserved keywords
Many keywords from SQL are now reserved keywords for RAQL queries. Column or function names that match reserved keywords must be delimited to prevent errors when queries are run.
See Handle Column or Function Name Incompatibilities for details and migration techniques.
Aggregate functions
It was possible in the previous release to include columns in the Select clause that were not used for grouping or for aggregate calculations when the query had a Group By clause or performed aggregation over the entire dataset. This is no longer valid.
Note:  
This does not apply to the use of aggregate functions with partitions or windows.
Aliases in the Select clause
It was possible in the previous release to refer to an alias defined earlier in the Select clause within the definition for a subsequent column. This is no longer valid.
No schema information
In some cases, errors can occur when queries involve datasets that do not have schema information to define their structure and datatypes. See Handle Errors from Derived Schemas for more information.
Use of quote marks (") versus apostrophes (')
Quote marks and apostrophes are no longer interchangeable:
*Quote marks are used to delimit names in query content that conflict with reserved words or that contain invalid characters.
*Apostrophes are used to delimit literal values.
Small changes in valid function names
Function names should not include either periods (.) or dashes (-).
A single period is supported as the delimiter between the name of the library that the function belongs to and the name of the function. For example:
myUDFLibrary.my_first_function
You can include periods or dashes in function names if you enclose their name in quote marks. For example:
myUDFLibrary."my-first-function"
Small changes in valid column names
Column names should not contain periods (.) or dashes (-). You can, however, enclose invalid names in quote marks to handle this limitation.
See the technique described previously for function names for details.
Case sensitivity for dataset variables, column names and function names.
In queries, the variable names that identify a dataset are case sensitive just as any variable name is in EMML. However, column and function names are not case sensitive.
Copyright © 2013-2015 Software AG, Darmstadt, Germany.

Product LogoContact Support   |   Community   |   Feedback