Hari Sekhon created DRILL-3529:
----------------------------------
Summary: Drill proper DESCRIBE support for CSV
Key: DRILL-3529
URL: https://issues.apache.org/jira/browse/DRILL-3529
Project: Apache Drill
Issue Type: Bug
Components: Metadata, Storage - Text & CSV
Affects Versions: 1.1.0
Reporter: Hari Sekhon
Assignee: Steven Phillips
Request to add full DESCRIBE support for CSV files.
Currently the describe command results in a blank table being printed instead
of the CSV header / schema.
This is dependent on DRILL-624 actually reading the header line as the schema
of the CSV.
After DRILL-624 is completed, I propose the following solution:
When dealing with a directory with multiple CSV files it would might make sense
to read N number of CSV file headers by default. Extend the DESCRIBE command to
have a user-configurable number of CSV file headers be read and presented, as
well as an ALL keywords to scan all CSV file headers to be able to
authoritatively print the schema of all the data.
It might also make sense to read the newest and oldest CSV files by timestamp
or by file name formatting conventions to pick up the different headers.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
