Github user xuchuanyin commented on the pull request:
https://github.com/apache/carbondata/commit/79feac96ae789851c5ad7306a7acaaba25d8e6c9#commitcomment-23407541
In
integration/spark2/src/main/scala/org/apache/spark/sql/hive/CarbonFileMetastore.scala:
In
integration/spark2/src/main/scala/org/apache/spark/sql/hive/CarbonFileMetastore.scala
on line 481:
@jackylk @sraghunandan Hi, Iâm studying the query related code and find
this line of code may cause a problem.
The member variable `metadata.tablesMeta` in `CarbonFileMetaStore` is a
list based, not Guava-Cache-like based. The action in `refreshCache` will
emptied the list but there is no corresponding action to fulfill it.
For example, in `CarbonSessionState.refreshRelationFromCache(...)`, it
firstly calls `checkSchemasModifiedTimeAndReloadTables`, which calls
`refreshCache` internally.
And then it will call `getTableFromMetadataCache`, which just does a search
in `metadata.tablesMeta`. But the value of `metadata.tablesMeta` is empty.
As a result, `CarbonSessionState.refreshRelationFromCache` actually does
nothing.
I haven't test it due to lack of environment. Could you review this and
correct me if I am wrong.
Ps: Keeping the `loadMetaData(...)` here is OK in my environment.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
