[ https://issues.apache.org/jira/browse/KYLIN-5635?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Xiaoxiang Yu updated KYLIN-5635: -------------------------------- Fix Version/s: (was: 5.0-beta) > Adapt for delta table > --------------------- > > Key: KYLIN-5635 > URL: https://issues.apache.org/jira/browse/KYLIN-5635 > Project: Kylin > Issue Type: Improvement > Components: RDBMS Source > Affects Versions: 5.0-alpha > Reporter: Zhiting Guo > Assignee: Zhiting Guo > Priority: Major > > When obtaining delta column information, because the column information is > not saved in the catalog (due to the use of DeltaCatalog), it cannot be > obtained directly from the catalog. We first determine whether the table is a > delta table. The Delta SDK provides a function. If so, read the table through > spark.table to get the schema. Here, spark scans Metadata to get the schema > information under the back path > > Also due to the use of DeltaCatalog, delta table does not support the show > create table statement, this is because deltaCatalog does some checks, does > not support this SQL , here by judging whether it is delta in advance, if it > is directly through location and table spell a ddl return. > > Limit: The partition column is not processed here, so the partition column is > not recognized, and delta does not manage its own partition through the > catalog. It is obtained in real time by scanning the Metadata under the > confidant path, so it does not affect the reading of data. The only place > that has an impact is the function of snapshot partition construction. -- This message was sent by Atlassian Jira (v8.20.10#820010)