[SPARK-7943] [SPARK-8105] [SPARK-8435] [SPARK-8714] [SPARK-8561] Fixes multi-database support #7623
Conversation
|
Test build #38244 has finished for PR 7623 at commit
|
|
|
||
| case _ => | ||
| val cmd = | ||
| CreateTableUsingAsSelect( |
There was a problem hiding this comment.
Should refactor CreateTableUsingAsSelect to use TableIdentifier.
|
Test build #38246 has finished for PR 7623 at commit
|
| table(new SqlParser().parseTableIdentifier(tableName)) | ||
| } | ||
|
|
||
| private[sql] def table(tableIdent: TableIdentifier): DataFrame = { |
There was a problem hiding this comment.
Is this called frequently enough to justify a function? It would be great to remove this if possible, since this API is also accessible in Java, and Java users won't see this as a private API.
There was a problem hiding this comment.
It's called right above... I can make it private instead of private[sql] though.
There was a problem hiding this comment.
I saw the call above. Why having this if it is only used once?
There was a problem hiding this comment.
To provide better support for multiple databases, internally we should always use TableIdentifier instead of String or Seq[String]. That's why I made this private[sql].
liancheng
force-pushed
the
spark-8131-multi-db
branch
from
c49cb5c
to
ec5f903
Compare
|
Test build #38460 has finished for PR 7623 at commit
|
|
Test build #38459 has finished for PR 7623 at commit
|
liancheng
force-pushed
the
spark-8131-multi-db
branch
from
ec5f903
to
f3bcd4b
Compare
|
Test build #38527 has finished for PR 7623 at commit
|
|
Merging to master. |
This PR fixes a set of issues related to multi-database. A new data structure
TableIdentifieris introduced to identify a table among multiple databases. We should stop using a singleString(table name without database name), orSeq[String](optional database name plus table name) to identify tables internally.