Thank you James.
I asked this b/c if the location is local file system the command show
tables seems work.
apache drill> use dfs.pyh;
+------+-------------------------------------+
| ok | summary |
+------+-------------------------------------+
| true | Default schema changed to [dfs.pyh] |
+------+-------------------------------------+
1 row selected (0.664 seconds)
apache drill (dfs.pyh)> show tables;
+--------------+------------+
| TABLE_SCHEMA | TABLE_NAME |
+--------------+------------+
| dfs.pyh | pplview |
| dfs.pyh | tt1 |
| dfs.pyh | tt2 |
| dfs.pyh | spamIP |
| dfs.pyh | ttt |
| dfs.pyh | foodmart |
+--------------+------------+
And yes 'show files' does work on HDFS path.
On 2022/4/6 12:07, James Turton wrote:
In filesystem storage a dataset will only be reported by SHOW FILES
unless you enable the Drill metastore and collect metadata from it using
ANALYZE TABLE ... REFRESH METADATA.
On 2022/04/06 03:42, Wes Peng wrote:
Hello the list,
Today I created a table in HDFS space:
apache drill (hdfs.tmp)> create table tt1 as select * from
`people.csvh` limit 1000;
I can run query to this table:
apache drill (hdfs.tmp)> select count(*) from tt1;
+--------+
| EXPR$0 |
+--------+
| 1000 |
+--------+
1 row selected (0.217 seconds)
And the table located in HDFS by listing the directory:
$ hdfs dfs -ls /tmp/test/tt1
Found 1 items
-rw-rw-r-- 3 pyh supergroup 100618 2022-04-06 09:37
/tmp/test/tt1/0_0_0.parquet
But the `show table` command does not work:
apache drill (hdfs.tmp)> show tables;
+--------------+------------+
| TABLE_SCHEMA | TABLE_NAME |
+--------------+------------+
No rows selected (0.177 seconds)
Do you know why?
Thanks.
