Impala Parquet 表分区表

最新推荐文章于 2025-09-16 22:20:10 发布

原创最新推荐文章于 2025-09-16 22:20:10 发布 · 3.1k 阅读

1 ·

CC 4.0 BY-SA版权

impala 专栏收录该内容

1 篇文章

订阅专栏

本文介绍如何使用 Impala 创建 Parquet 分区表并进行数据插入与查询操作，对比了 Parquet 分区表与普通表在数据加载速度上的表现。

1、Impala Parquet 表分区表

#parquet表
[hadoop02:21000] > create table parquet_ex_sms_event LIKE i_bil_hb_d.ex_sms_event stored as parquet;
Query: create table parquet_ex_sms_event LIKE i_bil_hb_d.ex_sms_event stored as parquet

[hadoop02:21000] > insert overwrite table parquet_ex_sms_event partition(date_no,area_code) select * from i_bil_hb_d.ex_sms_event where date_no=20141212 and area_code=0691;
Query: insert overwrite table parquet_ex_sms_event partition(date_no,area_code) select * from i_bil_hb_d.ex_sms_event where date_no=20141212 and area_code=0691

[hadoop02:21000] > select count(*) from parquet_ex_sms_event;
Query: select count(*) from parquet_ex_sms_event
+----------+
| count(*) |
+----------+
| 278678 |
+----------+
Fetched 1 row(s) in 0.58s

#普通表
[hadoop02:21000] > create table default_ex_sms_event as select * from i_bil_hb_d.ex_sms_event where date_no=20141212 and area_code=0691;
Query: create table default_ex_sms_event as select * from i_bil_hb_d.ex_sms_event where date_no=20141212 and area_code=0691
+------------------------+
| summary |
+------------------------+
| Inserted 278678 row(s) |
+------------------------+
Fetched 1 row(s) in 6.19s
[hadoop02:21000] > select count(*) from default_ex_sms_event;
Query: select count(*) from default_ex_sms_event
+----------+
| count(*) |
+----------+
| 278678 |
+----------+
Fetched 1 row(s) in 0.61s