DATE is not supported in CDH 5.16 Impala

[jpoblete]

Running impala-external.sql & impala-parquet.sql scripts on CDH 5.16.2 (impalad version 2.12.0-cdh5.16.2 RELEASE) fails with:

ERROR: AnalysisException: Unsupported data type: DATE

[jpoblete]

Had to change the scripts to execute...

sed -e 's/ date,/ varchar(10),/g' impala-external.sql  > impala-external_varchar.sql 
sed -e 's/ date,/ varchar(10),/g' impala-parquet.sql   > impala-parquet_varchar.sql

Also, the compute-stats.sql is missing:

use use tpcds_10000_parquet;

[drorke]

The DDL scripts were changed recently to use the DATE type which is only supported in Impala version 3.3 and later (https://issues.apache.org/jira/browse/IMPALA-7368). If you're looking for a Cloudera release with the required DATE support I believe it's supported in all of the CDP releases (both CDP cloud and CDP Data Center 7.0 for on-prem). The DATE support isn't available on CDH 5 or CDH 6. If you need to run on earlier CDH releases you could consider using a version of the TPC-DS kit prior to the recent commit that made the DATE related changes.

[gregrahn]

@jpoblete - stats are collected in impala-insert.sql -- compute-stats.sql is just a stand alone helper file if needed.