Zoltán Borók-Nagy created HIVE-21824:
----------------------------------------
Summary: Weird output format of 'describe formatted' command
Key: HIVE-21824
URL: https://issues.apache.org/jira/browse/HIVE-21824
Project: Hive
Issue Type: Bug
Affects Versions: 3.1.0
Reporter: Zoltán Borók-Nagy
Hive 3 changed the output format of the 'describe formatted' command. The new
format is vertical instead of the old horizontal format.
The expected output can be looked at the tests, e.g.:
[https://github.com/apache/hive/blob/10b6d70da1442cccf533bc97f56a622ec9f39661/ql/src/test/results/clientpositive/describe_table.q.out#L201-L216]
However, for Hive 3.1.0 (and probably for other Hive versions as well) the
output looks really strange:
{noformat}
jdbc:hive2://localhost:11050/default> describe formatted test i;
+------------------------+----------------------------------------------------+-------+-------+------------+-----------------+--------------+--------------+------------+-------------+------------+----------+
| col_name | data_type | min | max | num_nulls | distinct_count | avg_col_len |
max_col_len | num_trues | num_falses | bitvector | comment |
+------------------------+----------------------------------------------------+-------+-------+------------+-----------------+--------------+--------------+------------+-------------+------------+----------+
| col_name | i | NULL | NULL | NULL | NULL | NULL | NULL | NULL | NULL | NULL |
NULL |
| data_type | int | NULL | NULL | NULL | NULL | NULL | NULL | NULL | NULL |
NULL | NULL |
| min | 1 | NULL | NULL | NULL | NULL | NULL | NULL | NULL | NULL | NULL | NULL
|
| max | 3 | NULL | NULL | NULL | NULL | NULL | NULL | NULL | NULL | NULL | NULL
|
| num_nulls | 0 | NULL | NULL | NULL | NULL | NULL | NULL | NULL | NULL | NULL
| NULL |
| distinct_count | 3 | NULL | NULL | NULL | NULL | NULL | NULL | NULL | NULL |
NULL | NULL |
| avg_col_len | | NULL | NULL | NULL | NULL | NULL | NULL | NULL | NULL | NULL
| NULL |
| max_col_len | | NULL | NULL | NULL | NULL | NULL | NULL | NULL | NULL | NULL
| NULL |
| num_trues | | NULL | NULL | NULL | NULL | NULL | NULL | NULL | NULL | NULL |
NULL |
| num_falses | | NULL | NULL | NULL | NULL | NULL | NULL | NULL | NULL | NULL |
NULL |
| bitVector | | NULL | NULL | NULL | NULL | NULL | NULL | NULL | NULL | NULL |
NULL |
| comment | from deserializer | NULL | NULL | NULL | NULL | NULL | NULL | NULL
| NULL | NULL | NULL |
| COLUMN_STATS_ACCURATE |
{\"BASIC_STATS\":\"true\",\"COLUMN_STATS\":{\"i\":\"true\"}} | NULL | NULL |
NULL | NULL | NULL | NULL | NULL | NULL | NULL | NULL |
+------------------------+----------------------------------------------------+-------+-------+------------+-----------------+--------------+--------------+------------+-------------+------------+----------+{noformat}
Notice that the attribute names in the first column are the same as the column
names. And all the attribute values are listed in the 'data_type' column.
On cwiki there is also another example to this weird output at
[https://cwiki.apache.org/confluence/display/Hive/StatsDev#StatsDev-ExistingTables],
search for "the output would look like this". But it doesn't seem intentional
to me.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
