Can not query struct field with hive (CDH 5.9.0) - hadoop

Question

I just switch to CDH 5.9.0 (a full new install, not an upgrade, on a new cluster).
I have a table like this one (a bit more complex, but I reproduce with this example too):
CREATE TABLE `products`(`header` struct<PCODE:string, PNAME:string>)
PARTITIONED BY (`IMPORT_DATE' string)
ROW FORMAT SERDE
'org.apache.hadoop.hive.ql.io.parquet.serde.ParquetHiveSerDe'
STORED AS INPUTFORMAT
'org.apache.hadoop.hive.ql.io.parquet.MapredParquetInputFormat'
OUTPUTFORMAT
'org.apache.hadoop.hive.ql.io.parquet.MapredParquetOutputFormat'
LOCATION
'hdfs://myhost.com:8020/user/hive/warehouse/dbp/products'
TBLPROPERTIES ('transient_lastDdlTime'='1482160314')
If I do:
SELECT header FROM products;
==> The query is successful and return all products headers (in a JSON format)
But if I do:
SELECT header.PCODE FROM products;
==> It fails with the following stacktrace:
Error: java.lang.RuntimeException: Error in configuring object
at org.apache.hadoop.util.ReflectionUtils.setJobConf(ReflectionUtils.java:109)
at org.apache.hadoop.util.ReflectionUtils.setConf(ReflectionUtils.java:75)
at org.apache.hadoop.util.ReflectionUtils.newInstance(ReflectionUtils.java:133)
at org.apache.hadoop.mapred.MapTask.runOldMapper(MapTask.java:449)
at org.apache.hadoop.mapred.MapTask.run(MapTask.java:343)
at org.apache.hadoop.mapred.YarnChild$2.run(YarnChild.java:164)
at java.security.AccessController.doPrivileged(Native Method)
at javax.security.auth.Subject.doAs(Subject.java:422)
at org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1698)
at org.apache.hadoop.mapred.YarnChild.main(YarnChild.java:158)
Caused by: java.lang.reflect.InvocationTargetException
at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
at java.lang.reflect.Method.invoke(Method.java:498)
at org.apache.hadoop.util.ReflectionUtils.setJobConf(ReflectionUtils.java:106)
... 9 more
Caused by: java.lang.RuntimeException: Error in configuring object
at org.apache.hadoop.util.ReflectionUtils.setJobConf(ReflectionUtils.java:109)
at org.apache.hadoop.util.ReflectionUtils.setConf(ReflectionUtils.java:75)
at org.apache.hadoop.util.ReflectionUtils.newInstance(ReflectionUtils.java:133)
at org.apache.hadoop.mapred.MapRunner.configure(MapRunner.java:38)
... 14 more
Caused by: java.lang.reflect.InvocationTargetException
at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
at java.lang.reflect.Method.invoke(Method.java:498)
at org.apache.hadoop.util.ReflectionUtils.setJobConf(ReflectionUtils.java:106)
... 17 more
Caused by: java.lang.RuntimeException: Map operator initialization failed
at org.apache.hadoop.hive.ql.exec.mr.ExecMapper.configure(ExecMapper.java:147)
... 22 more
Caused by: java.lang.NullPointerException
at org.apache.hadoop.hive.ql.exec.ExprNodeFieldEvaluator.initialize(ExprNodeFieldEvaluator.java:61)
at org.apache.hadoop.hive.ql.exec.ExprNodeFieldEvaluator.initialize(ExprNodeFieldEvaluator.java:53)
at org.apache.hadoop.hive.ql.exec.Operator.initEvaluators(Operator.java:954)
at org.apache.hadoop.hive.ql.exec.Operator.initEvaluatorsAndReturnStruct(Operator.java:980)
at org.apache.hadoop.hive.ql.exec.SelectOperator.initializeOp(SelectOperator.java:63)
at org.apache.hadoop.hive.ql.exec.Operator.initialize(Operator.java:385)
at org.apache.hadoop.hive.ql.exec.Operator.initialize(Operator.java:469)
at org.apache.hadoop.hive.ql.exec.Operator.initializeChildren(Operator.java:425)
at org.apache.hadoop.hive.ql.exec.TableScanOperator.initializeOp(TableScanOperator.java:193)
at org.apache.hadoop.hive.ql.exec.Operator.initialize(Operator.java:385)
at org.apache.hadoop.hive.ql.exec.MapOperator.initializeOp(MapOperator.java:431)
at org.apache.hadoop.hive.ql.exec.Operator.initialize(Operator.java:385)
at org.apache.hadoop.hive.ql.exec.mr.ExecMapper.configure(ExecMapper.java:126)
... 22 more
On my old cluster (CDH 5.8.2), it works fine.
Any idea?
[EDIT: I have downgraded all CDH 5.9.0 jars (/opt/cloudera/parcels/CDH/jars) to the CDH 5.8.2 ones and the query is successfull. there is maybe a regression in CDH 5.9.0...]
[EDIT 2: If the table is stored as TextFile ( 'org.apache.hadoop.mapred.TextInputFormat'), the query runs successfuly.
We can think that the problem is linked with parquet.]
[Also posted on Cloudera forum: https://community.cloudera.com/t5/Batch-SQL-Apache-Hive/Can-not-query-struct-field-with-hive-CDH-5-9-0/m-p/48672#U48672 ]

Answer 1

I'll fix this by lowering the case on the query elements. P.ex:
SELECT header.pcode FROM products;

Answer 2

So I try a lot of things and I ended up with the followings:
-- Struct fieldnames in lowercase
CREATE TABLE `products`(`header` struct<pcode:string, pname:string>) STORED AS PARQUET;
Select results:
SELECT header.pcode FROM products ==> OK
SELECT HEADER.pcode FROM products ==> OK
SELECT header.PCODE FROM products ==> KO
SELECT HEADER.PCODE FROM products ==> KO
-- Struct fieldnames in UPPERCASE
CREATE TABLE `products`(`header` struct<PCODE:string, PNAME:string>) STORED AS PARQUET;
Select results:
SELECT header.pcode FROM products ==> KO
SELECT HEADER.pcode FROM products ==> KO
SELECT header.PCODE FROM products ==> KO
SELECT HEADER.PCODE FROM products ==> KO
==> Avoid UPPERCASE in struct fieldnames with tables stored as PARQUET in CDH 5.9.0 (it worked in CDH 5.8.2)...