[HIVE-1444] "hdfs" is hardcoded in few places in the code which inhibits use of other file systems

• Type: Bug
• Status: Closed
• Priority: Critical
• Resolution: Fixed
• Affects Version/s: 0.3.0, 0.4.0, 0.4.1, 0.5.0, 0.6.0, 0.7.0
• Fix Version/s: 0.9.0
• Component/s: Query Processor

Environment

any

• Assignee: Edward Capriolo
• Reporter: Yuliya Feldman
• Created: 30/Jun/10 21:26
• Updated: 09/Jan/13 10:24
• Resolved: 07/Mar/12 00:33

Description

In quite a few places "hdfs" is hardcoded, which is OK for majority of the cases, except when it is not really hdfs, but s3 or any other file system.

The place where it really breaks is:
in ql/src/java/org/apache/hadoop/hive/ql/parse/LoadSemanticAnalyzer.java :

method: private void applyConstraints(URI fromURI, URI toURI, Tree ast, boolean isLocal)

First few lines are check for file system:
if (!fromURI.getScheme().equals("file")
&& !fromURI.getScheme().equals("hdfs"))

{ throw new SemanticException(ErrorMsg.INVALID_PATH.getMsg(ast, "only \"file\" or \"hdfs\" file systems accepted")); }

"hdfs" is hardcoded.

I don't think you need to have this check at all as you are checking whether filesystem is local or not later on anyway and in regards to non locla file system - if one would be bad one you would get problems or have it look like local before you even come to "applyConstraints" method.