ravipesala commented on a change in pull request #3198: [CARBONDATA-3369] Fix
issues during concurrent execution of Create table If not exists
URL: https://github.com/apache/carbondata/pull/3198#discussion_r281472081
##########
File path:
integration/spark2/src/main/scala/org/apache/spark/sql/execution/command/table/CarbonCreateTableCommand.scala
##########
@@ -166,7 +167,35 @@ case class CarbonCreateTableCommand(
""".stripMargin)
}
} catch {
- case e: AnalysisException => throw e
+ case e: AnalysisException =>
+ // AnalysisException thrown with table already exists msg incase
of conurrent drivers
+ if (e.getMessage().contains("already exists")) {
+
+ // Clear the cache first
+ CarbonEnv.getInstance(sparkSession).carbonMetaStore
+ .removeTableFromMetadata(dbName, tableName)
+
+ // Delete the folders created by this call if the actual path is
different
+ val actualPath = CarbonEnv
+ .getCarbonTable(TableIdentifier(tableName,
Option(dbName)))(sparkSession)
+ .getTablePath
+
+ if (!actualPath.equalsIgnoreCase(tablePath)) {
+ LOGGER
+ .info(
+ "TableAlreadyExists with path : " + actualPath + " So,
deleting " + tablePath)
+
FileFactory.deleteAllCarbonFilesOfDir(FileFactory.getCarbonFile(tablePath))
+ }
+
+ // No need to throw for create if not exists
+ if (!ifNotExistsSet) {
Review comment:
But in old code, this check was not handled. where it was handled in old
code.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services