[CARBONDATA-3068] cannot load data from hdfs files without hdfs prefix

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 1.5.0
• Component/s: spark-integration
• Assignee: lianganping
• Reporter: lianganping
• Due: 02/Nov/18
• Created: 01/Nov/18 08:41
• Updated: 02/Nov/18 01:37
• Resolved: 02/Nov/18 01:37

Description

sql:
LOAD DATA INPATH '/tmp/test.csv' INTO TABLE test OPTIONS('QUOTECHAR'='"','TIMESTAMPFORMAT'='yyyy/MM/dd HH:mm:ss');
error:
org.apache.carbondata.processing.exception.DataLoadingException: The input file does not exist: /tmp/test.csv (state=,code=0)
but the file "test.csv" is in hdfs path, and hadoop conf "core-site.xml" has the property:

fs.defaultFS
hdfs://master:9000

Issue Links

• links to: GitHub Pull Request #2891