If the Hadoop property hive.exec.compress.output is set to TRUE, then jobs that leverage SAS/ACCESS Interface to Hadoop might stop responding or might return unexpected results. Unexpected results can include generating incorrect record counts or character fields that appear to be gibberish. SAS/ACCESS Interface to Hadoop jobs, which apply a WHERE statement or subsets columns, are most likely to see this issue since they generate a MapReduce to generate the output. Unexpected results might also be seen when the Hadoop engine is used in windowing environments such as SAS^® Enterprise Guide^® or the VIEWTABLE window to do these same types of activities.

Workaround

To circumvent this issue, do one of the following:

• Set hive.exec.compress.output to FALSE either by using PROPERTIES='hive.exec.compress.output=false' in the LIBNAME statement or by setting the property within the hive-site.xml file pointed to by the SAS_HADOOP_CONFIG_PATH environment variable.
• Use the READ_METHOD=JDBC option in the LIBNAME statement. This might result in slower Read performance.

Click the Hot Fix tab in this note to access the hot fix for this issue.

Operating System and Release Information

A fix for this issue for SAS/ACCESS Interface to Hadoop 9.43 is available at: