creating partition in external table in hive

问题

I have successfully created and added Dynamic partitions in an Internal table in hive. i.e. by using following steps:

1-created a source table

2-loaded data from local into source table

3- created another table with partitions - partition_table

4- inserted the data to this table from source table resulting in creation of all the partitions dynamically

My question is, how to perform this in external table? I read so many articles on this, but i am confused , that do I have to specify path to the already existing partitions for creating partitions for external table??

example: Step 1:

create external table1 ( name string, age int, height int)
location 'path/to/dataFile/in/HDFS';

Step 2:

alter table table1 add partition(age) 
location 'path/to/already/existing/partition'

I am not sure how to proceed with partitioning in external tables. Can somebody please help by giving step by step description of the same?.

Thanks in advance!

回答1:

1. Set below property

set hive.exec.dynamic.partition=true

set hive.exec.dynamic.partition.mode=nonstrict

2. Create External partitioned table

create external table1 ( name string, age int, height int) location 'path/to/dataFile/in/HDFS';

3. Insert data to partitioned table from source table.

Basically , the process is same. its just that you create external partitioned table and provide HDFS path to table under which it will create and store partition.

Hope this helps.

回答2:

Yes, you have to tell Hive explicitly what is your partition field.

Consider you have a following HDFS directory on which you want to create a external table.

/path/to/dataFile/

Let's say this directory already have data stored(partitioned) department wise as follows:

/path/to/dataFile/dept1
/path/to/dataFile/dept2
/path/to/dataFile/dept3

Each of these directories have bunch of files where each file contains actual comma separated data for fields say name,age,height.

e.g.
    /path/to/dataFile/dept1/file1.txt
    /path/to/dataFile/dept1/file2.txt

Now let's create external table on this:

Step 1. Create external table:

CREATE EXTERNAL TABLE testdb.table1(name string, age int, height int)
PARTITIONED BY (dept string)
ROW FORMAT DELIMITED
STORED AS TEXTFILE
LOCATION '/path/to/dataFile/';

Step 2. Add partitions:

ALTER TABLE testdb.table1 ADD PARTITION (dept='dept1') LOCATION '/path/to/dataFile/dept1';
ALTER TABLE testdb.table1 ADD PARTITION (dept='dept2') LOCATION '/path/to/dataFile/dept2';
ALTER TABLE testdb.table1 ADD PARTITION (dept='dept3') LOCATION '/path/to/dataFile/dept3';

Done, run select query once to verify if data loaded successfully.

回答3:

Follow the below steps:

1. Create a temporary table/Source table

   create table source_table(name string,age int,height int) row format delimited by ',';
   

   Use your delimiter as in the file instead of ',';

2. Load data into the source table

   load data local inpath 'path/to/dataFile/in/HDFS';
   

3. Create external table with partition

   create external table external_dynamic_partitions(name string,height int) 
   partitioned by (age int) 
   location 'path/to/dataFile/in/HDFS';
   

4. Enable dynamic partition mode to nonstrict

   set hive.exec.dynamic.partition.mode=nonstrict
   

5. Load data to external table with partitions from source file

   insert into table external_dynamic partition(age) 
   select * from source_table;
   

That's it. You can check the partitions information using

show partitions external_dynamic;

You can even check if it is an external table or not using

describe formatted external_dynamic;

External table is a type of table in Hive where the data is not moved to the hive warehouse. That means even if U delete the table, the data still persists and you will always get the latest data, which is not the case with Managed table.

来源：https://stackoverflow.com/questions/32580356/creating-partition-in-external-table-in-hive