上篇:第7章 函数
1、Hadoop源码编译支持Snappy压缩
1.1 资源准备
1.CentOS联网
配置CentOS能连接外网。Linux虚拟机ping www.baidu.com 是畅通的
注意:采用root角色编译,减少文件夹权限出现问题
2.jar包准备(hadoop源码、JDK8 、maven、protobuf)
(1)hadoop-2.7.2-src.tar.gz
(2)jdk-8u144-linux-x64.tar.gz
(3)snappy-1.1.3.tar.gz
(4)apache-maven-3.0.5-bin.tar.gz
(5)protobuf-2.5.0.tar.gz
1.2 jar包安装
注意:所有操作必须在root用户下完成
(1)JDK解压、配置环境变量JAVA_HOME和PATH,验证java-version(如下都需要验证是否配置成功)
[root@hadoop101 software] # tar -zxf jdk-8u144-linux-x64.tar.gz -C /opt/module/
[root@hadoop101 software]# vi /etc/profile
#JAVA_HOME
export JAVA_HOME=/opt/module/jdk1.8.0_144
export PATH=$PATH:$JAVA_HOME/bin
[root@hadoop101 software]#source /etc/profil
验证命令:java -version
(2)Maven解压、配置 MAVEN_HOME和PATH
[root@hadoop101 software]# tar -zxvf apache-maven-3.0.5-bin.tar.gz -C /opt/module/
[root@hadoop101 apache-maven-3.0.5]# vi /etc/profile
#MAVEN_HOME
export MAVEN_HOME=/opt/module/apache-maven-3.0.5
export PATH=$PATH:$MAVEN_HOME/bin
[root@hadoop101 software]#source /etc/profile
验证命令:mvn -version
1.3、检查hadoop本地库
[root@hadoop101 hadoop-2.7.2]# hadoop checknative
snappy: false
我们能查看到snappy不支持,所以需要把false改成true;
a、首先我们,需要把hadoop-2.7.2.tar.gz文件拷贝到/opt/module文件目录下:
解压该文件
[root@hadoop101 module]# tar -zxvf hadoop-2.7.2.tar.gz
刚解压这个文件后,查看这个文件信息
[root@hadoop101 module]# cd hadoop-2.7.2/lib
[root@hadoop101 lib]# ll
total 0
drwxr-xr-x 2 root root 285 Sep 1 2017 native
[root@hadoop101 lib]# cd native/
[root@hadoop101 native]# ll
total 5188
-rw-r--r-- 1 root root 1210260 Sep 1 2017 libhadoop.a
-rw-r--r-- 1 root root 1487268 Sep 1 2017 libhadooppipes.a
lrwxrwxrwx 1 root root 18 Sep 1 2017 libhadoop.so -> libhadoop.so.1.0.0
-rwxr-xr-x 1 root root 716316 Sep 1 2017 libhadoop.so.1.0.0
-rw-r--r-- 1 root root 582048 Sep 1 2017 libhadooputils.a
-rw-r--r-- 1 root root 364860 Sep 1 2017 libhdfs.a
lrwxrwxrwx 1 root root 16 Sep 1 2017 libhdfs.so -> libhdfs.so.0.0.0
-rwxr-xr-x 1 root root 229113 Sep 1 2017 libhdfs.so.0.0.0
-rw-r--r-- 1 root root 472950 Sep 1 2017 libsnappy.a
-rwxr-xr-x 1 root root 955 Sep 1 2017 libsnappy.la
lrwxrwxrwx 1 root root 18 Sep 1 2017 libsnappy.so -> libsnappy.so.1.3.0
lrwxrwxrwx 1 root root 18 Sep 1 2017 libsnappy.so.1 -> libsnappy.so.1.3.0
-rwxr-xr-x 1 root root 228177 Sep 1 2017 libsnappy.so.1.3.0
[root@hadoop101 native]#
以往的hadoop文件时支持的,如图所示:
不支持snappy:
所,我们需要做的就是,在可支持snappy文件把所有的文件拷贝到不支持文件下
[root@hadoop101 native]# ll
total 5188
-rw-r--r-- 1 root root 1210260 Sep 1 2017 libhadoop.a
-rw-r--r-- 1 root root 1487268 Sep 1 2017 libhadooppipes.a
lrwxrwxrwx 1 root root 18 Sep 1 2017 libhadoop.so -> libhadoop.so.1.0.0
-rwxr-xr-x 1 root root 716316 Sep 1 2017 libhadoop.so.1.0.0
-rw-r--r-- 1 root root 582048 Sep 1 2017 libhadooputils.a
-rw-r--r-- 1 root root 364860 Sep 1 2017 libhdfs.a
lrwxrwxrwx 1 root root 16 Sep 1 2017 libhdfs.so -> libhdfs.so.0.0.0
-rwxr-xr-x 1 root root 229113 Sep 1 2017 libhdfs.so.0.0.0
-rw-r--r-- 1 root root 472950 Sep 1 2017 libsnappy.a
-rwxr-xr-x 1 root root 955 Sep 1 2017 libsnappy.la
lrwxrwxrwx 1 root root 18 Sep 1 2017 libsnappy.so -> libsnappy.so.1.3.0
lrwxrwxrwx 1 root root 18 Sep 1 2017 libsnappy.so.1 -> libsnappy.so.1.3.0
-rwxr-xr-x 1 root root 228177 Sep 1 2017 libsnappy.so.1.3.0
[root@hadoop101 native]# cp libsnappy.a /usr/local/hadoop/module/hadoop-2.7.2/lib/native
覆盖,一直y,即可!
当前使用的文件已经拷贝覆盖进来了:
[root@hadoop101 native]# pwd
/usr/local/hadoop/module/hadoop-2.7.2/lib/native
[root@hadoop101 native]# ll
total 5636
-rw-r--r-- 1 root root 1210260 Jan 6 19:58 libhadoop.a
-rw-r--r-- 1 root root 1487268 Jan 6 20:04 libhadooppipes.a
lrwxrwxrwx 1 root root 18 May 22 2017 libhadoop.so -> libhadoop.so.1.0.0
-rwxr-xr-x 1 root root 716316 Jan 6 20:04 libhadoop.so.1.0.0
-rw-r--r-- 1 root root 582048 Jan 6 20:04 libhadooputils.a
-rw-r--r-- 1 root root 364860 Jan 6 20:04 libhdfs.a
lrwxrwxrwx 1 root root 16 May 22 2017 libhdfs.so -> libhdfs.so.0.0.0
-rwxr-xr-x 1 root root 229113 Jan 6 20:04 libhdfs.so.0.0.0
-rw-r--r-- 1 root root 472950 Jan 6 20:04 libsnappy.a
-rwxr-xr-x 1 root root 955 Jan 6 20:04 libsnappy.la
-rwxr-xr-x 1 root root 228177 Jan 6 20:04 libsnappy.so
-rwxr-xr-x 1 root root 228177 Jan 6 20:04 libsnappy.so.1
-rwxr-xr-x 1 root root 228177 Jan 6 20:04 libsnappy.so.1.3.0
[root@hadoop101 native]#
这时,我们需要关闭集群,重新启动集群
[root@hadoop101 native]# stop-all.sh
[root@hadoop101 native]# start-all.sh
这时,我们再次查看,hadoop本地库,发现支持为true
[root@hadoop101 native]# hadoop checknative
snappy: true /usr/local/hadoop/module/hadoop-2.7.2/lib/native/libsnappy.so.1
2、编译源码
(1)准备编译环境
[root@hadoop101 software]# yum install svn
[root@hadoop101 software]# yum install autoconf automake libtool cmake
[root@hadoop101 software]# yum install ncurses-devel
[root@hadoop101 software]# yum install openssl-devel
[root@hadoop101 software]# yum install gcc*
(2)编译安装snappy
[root@hadoop101 software]# tar -zxvf snappy-1.1.3.tar.gz -C /opt/module/
[root@hadoop101 module]# cd snappy-1.1.3/
[root@hadoop101 snappy-1.1.3]# ./configure
[root@hadoop101 snappy-1.1.3]# make
[root@hadoop101 snappy-1.1.3]# make install
# 查看snappy库文件
[root@hadoop101 snappy-1.1.3]# ls -lh /usr/local/lib |grep snappy
(3) 编译安装protobuf
[root@hadoop101 software]# tar -zxvf protobuf-2.5.0.tar.gz -C /opt/module/
[root@hadoop101 module]# cd protobuf-2.5.0/
[root@hadoop101 protobuf-2.5.0]# ./configure
[root@hadoop101 protobuf-2.5.0]# make
[root@hadoop101 protobuf-2.5.0]# make install
# 查看protobuf版本以测试是否安装成功
[root@hadoop101 protobuf-2.5.0]# protoc --version
(4)编译hadoop native
[root@hadoop101 software]# tar -zxvf hadoop-2.7.2-src.tar.gz
[root@hadoop101 software]# cd hadoop-2.7.2-src/
[root@hadoop101 software]# mvn clean package -DskipTests -Pdist,native -Dtar -Dsnappy.lib=/usr/local/lib -Dbundle.snappy
执行成功后,/opt/software/hadoop-2.7.2-src/hadoop-dist/target/hadoop-2.7.2.tar.gz即为新生成的支持snappy压缩的二进制安装包。
3、Hadoop压缩配置
3.1 MR支持的压缩编码
为了支持多种压缩/解压缩算法,Hadoop引入了编码/解码器,如下表所示:
压缩性能的比较:
http://google.github.io/snappy/
On a single core of a Core i7 processor in 64-bit mode, Snappy compresses at about 250 MB/sec or more and decompresses at about 500 MB/sec or more.
3.2 压缩参数配置
要在Hadoop中启用压缩,可以配置如下参数(mapred-site.xml文件中):
4、开启Map输出阶段压缩
开启map输出阶段压缩可以减少job中map和Reduce task间数据传输量。具体配置如下:
案例实操:
(1)开启hive中间传输数据压缩功能(启动:hive、beeline服务)
[root@hadoop101 ~]# cd /usr/local/hadoop/module/hive-1.2.1/bin/
[root@hadoop101 bin]# ./beeline
Beeline version 1.2.1 by Apache Hive
beeline> !connect jdbc:hive2://localhost:10000
Connecting to jdbc:hive2://localhost:10000
Enter username for jdbc:hive2://localhost:10000: root
Enter password for jdbc:hive2://localhost:10000:
Connected to: Apache Hive (version 1.2.1)
Driver: Hive JDBC (version 1.2.1)
Transaction isolation: TRANSACTION_REPEATABLE_READ
0: jdbc:hive2://localhost:10000>
0: jdbc:hive2://localhost:10000>
hadoop默认是:false,方式如下:
0: jdbc:hive2://localhost:10000> set hive.exec.compress.intermediate;
+----------------------------------------+--+
| set |
+----------------------------------------+--+
| hive.exec.compress.intermediate=false |
+----------------------------------------+--+
1 row selected (0.318 seconds)
0: jdbc:hive2://localhost:10000>
需要把它打开:
0: jdbc:hive2://localhost:10000> set hive.exec.compress.intermediate=true;
No rows affected (0.024 seconds)
0: jdbc:hive2://localhost:10000>
注意:需要指定它的编码格式:
0: jdbc:hive2://localhost:10000> set mapreduce.map.output.compress.codec;
+---------------------------------------------------------------------------------+--+
| set |
+---------------------------------------------------------------------------------+--+
| mapreduce.map.output.compress.codec=org.apache.hadoop.io.compress.DefaultCodec |
+---------------------------------------------------------------------------------+--+
1 row selected (0.02 seconds)
0: jdbc:hive2://localhost:10000>
设置mapreduce中map输出数据的压缩方式为:
0: jdbc:hive2://localhost:10000> set mapreduce.map.output.compress.codec=
0: jdbc:hive2://localhost:10000> org.apache.hadoop.io.compress.SnappyCodec;
No rows affected (0.004 seconds)
0: jdbc:hive2://localhost:10000>
接下来,我们尝试去执行查询语句
0: jdbc:hive2://localhost:10000> select count(ename) name from emp;
........
+-------+--+
| name |
+-------+--+
| 14 |
+-------+--+
1 row selected (39.854 seconds)
0: jdbc:hive2://localhost:10000>
5、开启Reduce输出阶段压缩
开启map输出阶段压缩可以减少job中map和Reduce task间数据传输量。具体配置如下:
案例实操:
(1)开启hive中间传输数据压缩功能
hive (default)> set hive.exec.compress.intermediate=true;
(2)开启mapreduce中map输出压缩功能
hive (default)>set hive.exec.compress.output=true;
(3)设置mapreduce最终数据输出压缩方式
hive (default)> set mapreduce.output.fileoutputformat.compress.codec =
org.apache.hadoop.io.compress.SnappyCodec;
(4)设置mapreduce最终数据输出压缩为块压缩
hive (default)> set mapreduce.output.fileoutputformat.compress.type=BLOCK;
(5)测试一下输出结果是否是压缩文件
hive (default)> insert overwrite local directory '/usr/local/hadoop/module/datas/distribute-result' select * from emp distribute by deptno sort by empno desc;
Query ID = root_20200106214406_879398ac-d7a1-4912-b421-e09eca1962d5
Total jobs = 1
Launching Job 1 out of 1
Number of reduce tasks not specified. Estimated from input data size: 1
In order to change the average load for a reducer (in bytes):
set hive.exec.reducers.bytes.per.reducer=<number>
In order to limit the maximum number of reducers:
set hive.exec.reducers.max=<number>
In order to set a constant number of reducers:
set mapreduce.job.reduces=<number>
Starting Job = job_1578341511756_0003, Tracking URL = http://hadoop101:8088/proxy/application_1578341511756_0003/
Kill Command = /usr/local/hadoop/module/hadoop-2.7.2/bin/hadoop job -kill job_1578341511756_0003
Hadoop job information for Stage-1: number of mappers: 1; number of reducers: 1
2020-01-06 21:44:25,101 Stage-1 map = 0%, reduce = 0%
2020-01-06 21:44:38,272 Stage-1 map = 100%, reduce = 0%, Cumulative CPU 1.82 sec
2020-01-06 21:44:48,365 Stage-1 map = 100%, reduce = 100%, Cumulative CPU 3.61 sec
MapReduce Total cumulative CPU time: 3 seconds 610 msec
Ended Job = job_1578341511756_0003
Copying data to local directory /usr/local/hadoop/module/datas/distribute-result
Copying data to local directory /usr/local/hadoop/module/datas/distribute-result
MapReduce Jobs Launched:
Stage-Stage-1: Map: 1 Reduce: 1 Cumulative CPU: 3.61 sec HDFS Read: 8330 HDFS Write: 661 SUCCESS
Total MapReduce CPU Time Spent: 3 seconds 610 msec
OK
emp.empno emp.ename emp.job emp.mgr emp.hiredate emp.sal emp.comm emp.deptno
Time taken: 44.329 seconds
hive (default)>
6、文件存储格式
Hive支持的存储数的格式主要有:TEXTFILE 、SEQUENCEFILE、ORC、PARQUET。
6.1 列式存储和行式存储
如图所示:左边为逻辑表,右边第一个为行式存储,第二个为列式存储。
(1)行存储的特点
查询满足条件的一整行数据的时候,列存储则需要去每个聚集的字段找到对应的每个列的值,行存储只需要找到其中一个值,其余的值都在相邻地方,所以此时行存储查询的速度更快。
(2)列存储的特点
因为每个字段的数据聚集存储,在查询只需要少数几个字段的时候,能大大减少读取的数据量;每个字段的数据类型一定是相同的,列式存储可以针对性的设计更好的设计压缩算法。
TEXTFILE和SEQUENCEFILE的存储格式都是基于行存储的;
ORC和PARQUET是基于列式存储的。
7、TextFile格式
默认格式,数据不做压缩,磁盘开销大,数据解析开销大。可结合Gzip、Bzip2使用,但使用Gzip这种方式,hive不会对数据进行切分,从而无法对数据进行并行操作。
Orc格式
Orc (Optimized Row Columnar)是Hive 0.11版里引入的新的存储格式。
如图6-11所示可以看到每个Orc文件由1个或多个stripe组成,每个stripe250MB大小,这个Stripe实际相当于RowGroup概念,不过大小由4MB->250MB,这样应该能提升顺序读的吞吐率。每个Stripe里有三部分组成,分别是Index Data,Row Data,Stripe Footer:
1)Index Data:一个轻量级的index,默认是每隔1W行做一个索引。这里做的索引应该只是记录某行的各字段在Row Data中的offset。
2)Row Data:存的是具体的数据,先取部分行,然后对这些行按列进行存储。对每个列进行了编码,分成多个Stream来存储。
3)Stripe Footer:存的是各个Stream的类型,长度等信息。
每个文件有一个File Footer,这里面存的是每个Stripe的行数,每个Column的数据类型信息等;每个文件的尾部是一个PostScript,这里面记录了整个文件的压缩类型以及FileFooter的长度信息等。在读取文件时,会seek到文件尾部读PostScript,从里面解析到File Footer长度,再读FileFooter,从里面解析到各个Stripe信息,再读各个Stripe,即从后往前读。
8、Parquet格式
Parquet是面向分析型业务的列式存储格式,由Twitter和Cloudera合作开发,2015年5月从Apache的孵化器里毕业成为Apache顶级项目。
Parquet文件是以二进制方式存储的,所以是不可以直接读取的,文件中包括该文件的数据和元数据,因此Parquet格式文件是自解析的。
通常情况下,在存储Parquet数据的时候会按照Block大小设置行组的大小,由于一般情况下每一个Mapper任务处理数据的最小单位是一个Block,这样可以把每一个行组由一个Mapper任务处理,增大任务执行并行度。
Parquet文件的格式如图
上图展示了一个Parquet文件的内容,一个文件中可以存储多个行组,文件的首位都是该文件的Magic Code,用于校验它是否是一个Parquet文件,Footer length记录了文件元数据的大小,通过该值和文件长度可以计算出元数据的偏移量,文件的元数据中包括每一个行组的元数据信息和该文件存储数据的Schema信息。除了文件中每一个行组的元数据,每一页的开始都会存储该页的元数据,在Parquet中,有三种类型的页:数据页、字典页和索引页。数据页用于存储当前行组中该列的值,字典页存储该列值的编码字典,每一个列块中最多包含一个字典页,索引页用来存储当前行组下该列的索引,目前Parquet中还不支持索引页。
9、主流文件存储格式对比实验
从存储文件的压缩比和查询速度两个角度对比。
存储文件的压缩比测试:
操作步骤
(1)首先把log.data文件上传到 /usr/local/hadoop/module/datas文件目录下:
(2)创建表,存储数据格式为TEXTFILE
2: jdbc:hive2://localhost:10000> create table log_text (
2: jdbc:hive2://localhost:10000> track_time string,
2: jdbc:hive2://localhost:10000> url string,
2: jdbc:hive2://localhost:10000> session_id string,
2: jdbc:hive2://localhost:10000> referer string,
2: jdbc:hive2://localhost:10000> ip string,
2: jdbc:hive2://localhost:10000> end_user_id string,
2: jdbc:hive2://localhost:10000> city_id string
2: jdbc:hive2://localhost:10000> )
2: jdbc:hive2://localhost:10000> row format delimited fields terminated by '\t'
2: jdbc:hive2://localhost:10000> stored as textfile ;
No rows affected (0.507 seconds)
2: jdbc:hive2://localhost:10000>
注意:创建数据表必须以root方式创建才行,不然会出错!
(3)向表中加载数据
2: jdbc:hive2://localhost:10000> load data local inpath '/usr/local/hadoop/module/datas/log.data'
2: jdbc:hive2://localhost:10000> into table log_text ;
INFO : Loading data to table default.log_text from file:/usr/local/hadoop/module/datas/log.data
INFO : Table default.log_text stats: [numFiles=1, totalSize=19014996]
No rows affected (3.312 seconds)
2: jdbc:hive2://localhost:10000>
在HDFS文件系统查看:数据加载进来了
(4)查看表中数据大小
2: jdbc:hive2://localhost:10000> dfs -du -h /user/hive/warehouse/log_text;
+-------------------------------------------------+--+
| DFS Output |
+-------------------------------------------------+--+
| 18.1 M /user/hive/warehouse/log_text/log.data |
+-------------------------------------------------+--+
1 row selected (0.109 seconds)
2: jdbc:hive2://localhost:10000>
9.1 ORC
(1)创建表,存储数据格式为ORC
2: jdbc:hive2://localhost:10000> create table log_orc(
2: jdbc:hive2://localhost:10000> track_time string,
2: jdbc:hive2://localhost:10000> url string,
2: jdbc:hive2://localhost:10000> session_id string,
2: jdbc:hive2://localhost:10000> referer string,
2: jdbc:hive2://localhost:10000> ip string,
2: jdbc:hive2://localhost:10000> end_user_id string,
2: jdbc:hive2://localhost:10000> city_id string
2: jdbc:hive2://localhost:10000> )
2: jdbc:hive2://localhost:10000> row format delimited fields terminated by '\t'
2: jdbc:hive2://localhost:10000> stored as orc ;
No rows affected (0.209 seconds)
No rows affected (0.273 seconds)
2: jdbc:hive2://localhost:10000>
(2)向表中加载数据
2: jdbc:hive2://localhost:10000> insert into table log_orc select * from log_text ;
在HDFS文件系统查看:数据加载进来了
(3) 查看插入后数据
2: jdbc:hive2://localhost:10000> dfs -du -h /user/hive/warehouse/log_orc/ ;
+-----------------------------------------------+--+
| DFS Output |
+-----------------------------------------------+--+
| 2.8 M /user/hive/warehouse/log_orc/000000_0 |
+-----------------------------------------------+--+
1 row selected (0.04 seconds)
2: jdbc:hive2://localhost:10000>
由此可见:创建出来的TEXTFILE文件格式与ORC文件格式所占的内存不一样:
TEXTFILE文件格式: 18.1 M
ORC文件格式:2.8 M
9.2 Parquet格式
(1)创建表,存储数据格式为parquet
1 row selected (0.04 seconds)
2: jdbc:hive2://localhost:10000> create table log_parquet(
2: jdbc:hive2://localhost:10000> track_time string,
2: jdbc:hive2://localhost:10000> url string,
2: jdbc:hive2://localhost:10000> session_id string,
2: jdbc:hive2://localhost:10000> referer string,
2: jdbc:hive2://localhost:10000> ip string,
2: jdbc:hive2://localhost:10000> end_user_id string,
2: jdbc:hive2://localhost:10000> city_id string
2: jdbc:hive2://localhost:10000> )
2: jdbc:hive2://localhost:10000> row format delimited fields terminated by '\t'
2: jdbc:hive2://localhost:10000> stored as parquet ;
(2)向表中加载数据
2: jdbc:hive2://localhost:10000> insert into table log_parquet select * from log_text ;
在HDFS文件系统查看:数据加载进来了
(3)查看表中数据大小
2: jdbc:hive2://localhost:10000> dfs -du -h /user/hive/warehouse/log_parquet/ ;
+----------------------------------------------------+--+
| DFS Output |
+----------------------------------------------------+--+
| 13.1 M /user/hive/warehouse/log_parquet/000000_0 |
+----------------------------------------------------+--+
1 row selected (0.071 seconds)
2: jdbc:hive2://localhost:10000>
(4)查看这张表使用是数据量与查询数据
log_parquet表:
2: jdbc:hive2://localhost:10000> select count(*)from log_parquet;
+---------+--+
| _c0 |
+---------+--+
| 100000 |
+---------+--+
1 row selected (46.892 seconds)
2: jdbc:hive2://localhost:10000>
有10万条数据,消耗了46S
log_orc表
2: jdbc:hive2://localhost:10000> select count(*)from log_orc;
+---------+--+
| _c0 |
+---------+--+
| 100000 |
+---------+--+
1 row selected (39.835 seconds)
有10万条数据,消耗了39S
0: jdbc:hive2://localhost:10000> select count(*)from log_text;
log_text数据表
+---------+--+
| _c0 |
+---------+--+
| 100000 |
+---------+--+
1 row selected (38.502 seconds)
有10万条数据,消耗了38S
在这几种方式之中,调优建议使用:orc
ORC与TEXTFILE结合使用
0: jdbc:hive2://localhost:10000> insert into table log_parquet select * from log_text ;
(1)创建一个非压缩的的ORC存储方式
0: jdbc:hive2://localhost:10000> create table log_orc_none(
0: jdbc:hive2://localhost:10000> track_time string,
0: jdbc:hive2://localhost:10000> url string,
0: jdbc:hive2://localhost:10000> session_id string,
0: jdbc:hive2://localhost:10000> referer string,
0: jdbc:hive2://localhost:10000> ip string,
0: jdbc:hive2://localhost:10000> end_user_id string,
0: jdbc:hive2://localhost:10000> city_id string
0: jdbc:hive2://localhost:10000> )
0: jdbc:hive2://localhost:10000> row format delimited fields terminated by '\t'
0: jdbc:hive2://localhost:10000> stored as orc tblproperties ("orc.compress"="NONE");
(2)插入数据
hive (default)> insert into table log_orc_none select * from log_text ;
(3)查看插入后数据
0: jdbc:hive2://localhost:10000> dfs -du -h /user/hive/warehouse/log_orc_none/ ;
+-----------------------------------------------------------+--+
| DFS Output |
+-----------------------------------------------------------+--+
| 7.7 M /user/hive/warehouse/log_orc_none/000000_0 |
| 7.7 M /user/hive/warehouse/log_orc_none/000000_0_copy_1 |
+-----------------------------------------------------------+--+
2 rows selected (0.055 seconds)
0: jdbc:hive2://localhost:10000>
2、创建一个SNAPPY压缩的ORC存储方式
(1)建表语句
0: jdbc:hive2://localhost:10000> create table log_orc_snappy(
0: jdbc:hive2://localhost:10000> track_time string,
0: jdbc:hive2://localhost:10000> url string,
0: jdbc:hive2://localhost:10000> session_id string,
0: jdbc:hive2://localhost:10000> referer string,
0: jdbc:hive2://localhost:10000> ip string,
0: jdbc:hive2://localhost:10000> end_user_id string,
0: jdbc:hive2://localhost:10000> city_id string
0: jdbc:hive2://localhost:10000> )
0: jdbc:hive2://localhost:10000> row format delimited fields terminated by '\t'
0: jdbc:hive2://localhost:10000> stored as orc tblproperties ("orc.compress"="SNAPPY");
No rows affected (0.257 seconds)
0: jdbc:hive2://localhost:10000>
(2) 插入数据
0: jdbc:hive2://localhost:10000> insert into table log_orc_snappy select * from log_text ;
(3)查看插入后数据
3、默认创建的ORC存储方式,导入数据后的大小为
2.8 M /user/hive/warehouse/log_orc/000000_0
比Snappy压缩的还小。原因是orc存储文件默认采用ZLIB压缩。比snappy压缩的小。
4.存储方式和压缩总结
在实际的项目开发当中,hive表的数据存储格式一般选择:orc或parquet。压缩方式一般选择snappy,lzo。
5、测试存储和压缩
官网:https://cwiki.apache.org/confluence/display/Hive/LanguageManual+ORC
ORC存储方式的压缩: