According to docs, this command should return table structure:
DESCRIBE schema.<table>
I tried:
-- Some CSV of 1000 lines (csv.extract_header=true)
DESCRIBE dfs.`sample_data.csv`
-- Empty result
-- Output format set to parquet
CREATE TABLE dfs.tmp.`test_parquet` AS SELECT * FROM dfs.`sample_data.csv`
-- > 1000 rows created
-- Querying newly created table
DESCRIBE dfs.tmp.`test_parquet`
-- Empty result
-- Querying my own parquet from somewhere else
DESCRIBE dfs.`another.parquet`
-- Empty result
REFRESH TABLE METADATA dfs.tmp.`test_parquet`
DESCRIBE dfs.tmp.`test_parquet`
-- Empty result
-- I've got postgres datasource connected as "pg"
DESCRIBE pg.public.MyPostgresTable
-- > Returns valid structure
It seems that DESCRIBE
command does not describe parquets and CSV.
What I'm missing?
Drill v1.15
According to docs, this command may be used for views created in a workspace, tables created in Hive and HBase, or schemas.
DESCRIBE
does not support tables created in a file system.