drill-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From HanumathRao <...@git.apache.org>
Subject [GitHub] drill pull request #996: DRILL-5878: TableNotFound exception is being report...
Date Sun, 22 Oct 2017 00:04:21 GMT
Github user HanumathRao commented on a diff in the pull request:

    https://github.com/apache/drill/pull/996#discussion_r146118431
  
    --- Diff: exec/java-exec/src/main/java/org/apache/drill/exec/planner/sql/SchemaUtilites.java
---
    @@ -77,6 +77,22 @@ public static SchemaPlus findSchema(final SchemaPlus defaultSchema,
final String
         return findSchema(defaultSchema, schemaPathAsList);
       }
     
    +  /**
    +   * Utility function to get the commonPrefix schema between two supplied schemas.
    +   * @param defaultSchema default schema
    +   * @param schemaPath current schema path
    +   * @param isCaseSensitive true if caseSensitive comparision is required.
    +   * @return common prefix schemaPath
    +   */
    +  public static String getPrefixSchemaPath(final String defaultSchema, final String schemaPath,
final boolean isCaseSensitive) {
    +    if (!isCaseSensitive) {
    +      return Strings.commonPrefix(defaultSchema.toLowerCase(), schemaPath.toLowerCase());
    --- End diff --
    
    @paul-rogers  Thank you for this catch. I did try out some examples which were throwing
schemanotfound exception. I have changed the code to handle it. getDefaultSchemaPath of session
object returns the defaultSchema as String, hence I converted the SchemaPath to String by
SchemaPathJOINER. As pointed out by your comment, this will loose some context information
if the "." is part of the name. However, I also think that there is a bug in the defaultSchema
code path as it is also loosing this information by converting it to String. For now I am
not fixing that bug (which I am planning to fix in another JIRA) as I am not sure   about
the changes in setting defaultSchema.
    
    Coming on to the specific valid use case which was throwing schemaNotFound exception is
    in dfs I have introduced "tmp.tmp1" in place of tmp.
    
    And if I query select * from dfs.`tmp.tmp1`.`employee.json` limit 1; this query reports
SchemaNotFound exception even though a workspace with name tmp.tmp1 exists.
    
    This is currently working with new changes.


---

Mime
View raw message