Hive: how to show all partitions of a table?

Question

I have a table with 1000+ partitions.

"Show partitions" command only lists a small number of partitions.

How can i show all partitions?

Update:

1. I found "show partitions" command only lists exactly 500 partitions.

2. "select ... where ..." only processes the 500 partitions!

Answer 1

CLI has some limit when ouput is displayed. I suggest to export output into local file:

$hive -e 'show partitions table;' > partitions

Answer 2

hive> show partitions table_name;

Answer 3

You can see Hive MetaStore tables,Partitions information in table of "PARTITIONS". You could use "TBLS" join "Partition" to query special table partitions.

Answer 4

Okay, I'm writing this answer by extending wmky's answer above & also, assuming that you've configured mysql for your metastore instead of derby.

select PART_NAME FROM PARTITIONS WHERE TBL_ID=(SELECT TBL_ID FROM TBLS WHERE TBL_NAME='<table_name>');

The above query gives you all possible values of the partition columns.

Example:

hive> desc clicks_fact;
OK
time                    timestamp                                   
..                              
day                     date                                        
file_date               varchar(8)                                  

# Partition Information      
# col_name              data_type               comment             

day                     date                                        
file_date               varchar(8)                                  
Time taken: 1.075 seconds, Fetched: 28 row(s)

I'm going to fetch the values of partition columns.

mysql> select PART_NAME FROM PARTITIONS WHERE TBL_ID=(SELECT TBL_ID FROM TBLS WHERE TBL_NAME='clicks_fact');
+-----------------------------------+
| PART_NAME                         |
+-----------------------------------+
| day=2016-08-16/file_date=20160816 |
| day=2016-08-17/file_date=20160816 |
....
....
| day=2017-09-09/file_date=20170909 |
| day=2017-09-08/file_date=20170909 |
| day=2017-09-09/file_date=20170910 |
| day=2017-09-10/file_date=20170910 |
+-----------------------------------+

1216 rows in set (0.00 sec)

Returns all partition columns.

Note: JOIN table DBS ON DB_ID when there is a DB involved (i.e, when, multiple DB's have same table_name)

Answer 5

Yet another option is to communicate with Hive Metastore via Thrift protocol.
If you write code in python, you may benefit from hmsclient library:

Hive cli:

hive> create table test_table_with_partitions(f1 string, f2 int) partitioned by (dt string);
OK
Time taken: 0.127 seconds

hive> alter table test_table_with_partitions add partition(dt=20210504) partition(dt=20210505);
OK
Time taken: 0.152 seconds

Python cli:

>>> from hmsclient import hmsclient
>>> client = hmsclient.HMSClient(host='hive.metastore.location', port=9083)
>>> with client as c:
...    all_partitions = c.get_partitions(db_name='default',
...                                      tbl_name='test_table_with_partitions', 
...                                      max_parts=24 * 365 * 3)
...
>>> print([{'dt': part.values[0]} for part in all_partitions])
[{'dt': '20210504'}, {'dt': '20210505'}]

NB: max_parts is a parameter than cannot be greater than 32767 (java short max value).

If you have Airflow installed together with apache.hive extra, you create hmsclient quite easy:

hive_hook = HiveMetastoreHook()
with hive_hook.metastore as hive_client:
    ... your code goes here ...

This seems a more valid way of communication with Hive Metastore than accessing DB directly (and database engine agnostic BTW).