[hadoop][hive]hive sort by and order by

[hadoop][hive]hive sort by and order by


Order by

• Toto ordering of query result set


• All data is passed through a single reducer

Sort by

• Order the data within each reducer ( local ordering) 


• Reducer’s output will be sorted

 

 

1. 使用 order by 在reduce 階段，因為全排序(也就是要整個排序完資料，才能做後續的處理)，所以會集中在一台reducer上面做操作。


2. 使用 sort by 會幫你做 local order 在做merge sort 。效率會比較高。

 

兩個語法使用範例如下:

SELECT s.ymd,s.name,s.price
FROM stocks s
ORDER BY s.ymd ASC ,s.name DESC;


SELECT s.ymd,s.name,s.price
FROM stocks s
SORT BY s.ymd ASC ,s.name DESC;