我在 HIVE 中有三个外部表:
表格1:
CREATE EXTERNAL TABLE IF NOT EXISTS table_1(
unique_key_column_1 VARCHAR,
column_needed_1 DATE,
redundant_column_1 VARCHAR,
redundant_column_2 VARCHAR,
redundant_column_3 VARCHAR,
column_needed_2 TIMESTAMP,
redundant_column_4 VARCHAR,
redundant_column_5 VARCHAR,
column_needed_3 INT,
redundant_column_6 VARCHAR,
redundant_column_7 VARCHAR)
ROW FORMAT DELIMITED FIELDS TERMINATED BY ',’
STORED AS TEXTFILE location '/user/<username>/visdata';
表2:
CREATE EXTERNAL TABLE IF NOT EXISTS table_2(
unique_key_column_1 VARCHAR,
column_needed_4 VARCHAR,
column_needed_5 VARCHAR,
unique_key_column_2 VARCHAR,
redundant_column_1 VARCHAR,
redundant_column_2 VARCHAR,
redundant_column_3 VARCHAR,
column_needed_6 TINYINT,
redundant_column_4 VARCHAR,
redundant_column_5 VARCHAR,
column_needed_7 DATE,
redundant_column_6 VARCHAR,
redundant_column_7 VARCHAR)
ROW FORMAT DELIMITED FIELDS TERMINATED BY ',’
STORED AS TEXTFILE location '/user/<username>/visdata';
表3:
CREATE EXTERNAL TABLE IF NOT EXISTS table_3(
unique_key_column_2 VARCHAR,
redundant_column_1 VARCHAR,
redundant_column_2 VARCHAR,
redundant_column_3 VARCHAR,
redundant_column_4 VARCHAR,
redundant_column_5 VARCHAR,
column_needed_8 VARCHAR,
column_needed_9 TINYINT,
redundant_column_6 VARCHAR,
redundant_column_7 VARCHAR,
column_needed_10 TIMESTAMP)
ROW FORMAT DELIMITED FIELDS TERMINATED BY ',’
STORED AS TEXTFILE location '/user/<username>/visdata';
我现在想创建一个托管表,在我的两个唯一键列上的表上方左外连接,如下所示:
unique_key_column_1 column_needed_1 column_needed_2 column_needed_3 column_needed_4 column_needed_5 column_needed_1 column_needed_6 column_needed_7 unique_key_column_2 column_needed_8 column_needed_9 column_needed_10
key_entry_1_1 entry_1_1 entry_1_2 entry_1_3 entry_1_4 entry_1_5 entry_1_6 entry_1_7 key_entry_1_2 entry_1_8 entry_1_9 entry_1_10
key_entry_2_2 entry_2_1 entry_2_2 entry_2_3 entry_2_4 entry_2_5 entry_2_6 entry_2_7 key_entry_2_2 entry_2_8 entry_2_9 entry_2_10
我该怎么做呢?
编辑1:
这是我能想到的,从两张桌子加入。我仍然无法弄清楚如何将三个表连接成一个表:
> create table combined_table;
> insert into combined_table SELECT * FROM (SELECT r.unique_key_column_1, r.column_needed_1, r.column_needed_2, r.column_needed_3, o.r.column_needed_4, o.column_needed_5, o.column_needed_6, o.column_needed_7 FROM table_1 LEFT OUTER JOIN table_2 o ON (r.unique_key_column_1 = o.unique_key_column_2 );
编辑2:
我刚刚意识到加入很昂贵。那么,有什么我可以使用分区来做到这一点的吗?