Dave Challis created DRILL-7732: ----------------------------------- Summary: Group by on Parquet field causing: SYSTEM ERROR: CompileException Key: DRILL-7732 URL: https://issues.apache.org/jira/browse/DRILL-7732 Project: Apache Drill Issue Type: Bug Affects Versions: 1.17.0 Environment: AWS Linux instance, 64Gb RAM. Reporter: Dave Challis Attachments: portfolio.parquet
When executing the following query in Drill (via the web UI): {noformat} SELECT P.portfolio_id FROM dfs.`/data/portfolio.parquet` AS P GROUP BY P.portfolio_id {noformat} the query fails with the following error message: {noformat} org.apache.drill.common.exceptions.UserRemoteException: SYSTEM ERROR: CompileException: Line 61, Column 30: "value" is neither a method, a field, nor a member class of "org.apache.drill.exec.vector.UntypedNullHolder" Fragment 0:0 Please, refer to logs for more information. [Error Id: be0b9712-c9c6-4733-9d17-5c370d5dfce0 on e531a6492cf4:31010] {noformat} The schema for the Parquet file in question is: {noformat} $ parquet-tool schema portfolio.parquet message parquet_go_root { required int32 portfolio_id (UINT_32) = 0; required binary name (UTF8) = 0; optional binary meta (UTF8) = 0; } {noformat} The contents of the file is a single row: {noformat} $ parquet-tools cat portfolio.parquet portfolio_id = 0 name = Bose Corporation meta = null {noformat} I've attached this file to this ticket. -- This message was sent by Atlassian Jira (v8.3.4#803005)