The following usage
notes apply to Cloudera Impala, a query engine that runs on Apache
Hadoop.
Impala does not support
SQL DELETE or UPDATE statements. Accordingly, any SAS Data Integration
Studio transformation that generates SQL DELETE or UPDATE statements
will fail if these operations are attempted on an Impala table. Affected
transformations include the SQL Delete transformation, the SQL Update
transformation, and the Table Loader transformation.
For more information
about the loader transformation, see Cloudera Impala: Usage Notes for Loaders.
Impala Requires SQL
ORDER BY statements to specify a limit. The SQL Join transformation
does not specify a limit on an ORDER BY statement by default. Accordingly,
if an SQL Join transformation attempts to execute an ORDER BY statement,
the transformation will fail. If explicit pass-through is set for
the SQL Join, one remedy is to add the limit clause in the generated
code.
One way to add the limit
clause is as follows:
-
Open the properties
window for the SQL Join.
-
-
Change the Code generation
mode to All user written.
-
Add the limit clause
after the ORDER BY clause.
Here is an example where
the limit is 100 observations:
order by <column> limit 100
-
Click
OK to
close the properties window and save your changes.
If implicit pass-through
is set for the SQL join, one remedy is to specify a data set option,
such as
OBS=100, in the properties
window for the Impala table.
One way to specify a
data set option is as follows:
-
Open the properties
window for the table.
-
Select
Options tabGeneralAdditional Table Options.
-
Specify an option such
as
OBS=100.