资讯

精准传达 • 有效沟通

从品牌网站建设到网络营销策划,从策略到执行的一站式服务

hbase的四种压缩格式

Hbase支持的压缩格式:

拜泉ssl适用于网站、小程序/APP、API接口等需要进行数据传输应用场景,ssl证书未来市场广阔!成为创新互联的ssl证书销售渠道,可以享受市场价格4-6折优惠!如果有意向欢迎电话联系或者加微信:13518219792(备注:SSL证书合作)期待与您的合作!

hbase支持的压缩格式:GZ(GZIP),LZ0,LZ4,Snappy

GZ:用于冷数据压缩,与Snappy和LZ0相比,GZIP的压缩率更高,但是更消耗CPU,解压/压缩速度更慢。

Snappy和LZ0:用于热数据压缩,占用CPU少,解压/压缩速度比GZ快,但是压缩率不如GZ高。

Snappy与LZ0相比,Snappy整体性能优于LZ0,Snappy压缩率比LZ0更低,但是解压/压缩速度更快。

LZ4与LZ0相比,LZ4的压缩率和LZ0的压缩率相差不多,但是LZ4的解压/压缩速度更快。

多数情况下,选择Snppy或LZ0是比较好的选择,因为它们的压缩开销底,能节省空间。

 

建表时指定压缩格式

hbase(main):013:0> create 'test3',{NAME=>'f1'},{NAME=>'f2',COMPRESSION=>'Snappy'}
0 row(s) in 1.2740 seconds
 
=> Hbase::Table - test3
hbase(main):014:0> desc 'test3'
Table test3 is ENABLED                                                                                                                                             
test3                                                                                                                                                              
COLUMN FAMILIES DESCRIPTION                                                                                                                                        
{NAME => 'f1', BLOOMFILTER => 'ROW',VERSIONS => '1', IN_MEMORY => 'false', KEEP_DELETED_CELLS => 'FALSE',DATA_BLOCK_ENCODING => 'NONE', TTL => 'FOREVER', COMPRESS
ION => 'NONE', MIN_VERSIONS => '0',BLOCKCACHE => 'true', BLOCKSIZE => '65536', REPLICATION_SCOPE =>'0'}                                                           
{NAME => 'f2', BLOOMFILTER => 'ROW',VERSIONS => '1', IN_MEMORY => 'false', KEEP_DELETED_CELLS => 'FALSE',DATA_BLOCK_ENCODING => 'NONE', TTL => 'FOREVER', COMPRESS
ION => 'SNAPPY', MIN_VERSIONS => '0',BLOCKCACHE => 'true', BLOCKSIZE => '65536', REPLICATION_SCOPE =>'0'}                                                        
2 row(s) in 0.0300 seconds

hbase(main):002:0> create 'test4' ,{NAME=>'f1'},{NAME=>'f2',COMPRESSION=>'GZ'}
0 row(s) in 1.4900 seconds

=> Hbase::Table - test4
hbase(main):003:0> desc 'test4'
Table test4 is ENABLED                                                                                                                                              
test4                                                                                                                                                               
COLUMN FAMILIES DESCRIPTION                                                                                                                                         
{NAME => 'f1', BLOOMFILTER => 'ROW', VERSIONS => '1', IN_MEMORY => 'false', KEEP_DELETED_CELLS => 'FALSE', DATA_BLOCK_ENCODING => 'NONE', TTL => 'FOREVER', COMPRESS
ION => 'NONE', MIN_VERSIONS => '0', BLOCKCACHE => 'true', BLOCKSIZE => '65536', REPLICATION_SCOPE => '0'}                                                           
{NAME => 'f2', BLOOMFILTER => 'ROW', VERSIONS => '1', IN_MEMORY => 'false', KEEP_DELETED_CELLS => 'FALSE', DATA_BLOCK_ENCODING => 'NONE', TTL => 'FOREVER', COMPRESS
ION => 'GZ', MIN_VERSIONS => '0', BLOCKCACHE => 'true', BLOCKSIZE => '65536', REPLICATION_SCOPE => '0'}                                                             
2 row(s) in 0.1290 seconds

 

建表后修改columnfamily压缩格式

正确做法是先disable表,再修改列族压缩格式,enbale表后做major_compact操作。

如下:

hbase(main):004:0> desc 'test1'
Table test1 is ENABLED                                                                                                                                             
test1                                                                                                                                                              
COLUMN FAMILIES DESCRIPTION                                                                                                                                        
{NAME => 'f1', BLOOMFILTER => 'ROW',VERSIONS => '1', IN_MEMORY => 'false', KEEP_DELETED_CELLS => 'FALSE',DATA_BLOCK_ENCODING => 'NONE', TTL => 'FOREVER', COMPRESS
ION => 'NONE', MIN_VERSIONS => '0',BLOCKCACHE => 'true', BLOCKSIZE => '65536', REPLICATION_SCOPE =>'0'}                                                          
{NAME => 'f2', BLOOMFILTER => 'ROW',VERSIONS => '1', IN_MEMORY => 'false', KEEP_DELETED_CELLS => 'FALSE',DATA_BLOCK_ENCODING => 'NONE', TTL => 'FOREVER', COMPRESS
ION => 'NONE', MIN_VERSIONS => '0',BLOCKCACHE => 'true', BLOCKSIZE => '65536', REPLICATION_SCOPE =>'0'}                                                          
2 row(s) in 0.0230 seconds
 
hbase(main):005:0> disable 'test1'
0 row(s) in 2.2870 seconds
 
hbase(main):006:0> alter 'test1' ,{NAME=>'f1',COMPRESSION=>'Snappy'}
Updating all regions with the new schema...
1/1 regions updated.
Done.
0 row(s) in 1.9510 seconds
 
hbase(main):007:0> enable 'test1'
0 row(s) in 1.2820 seconds
 
hbase(main):008:0> desc 'test1'
Table test1 is ENABLED                                                                                                                                              
test1                                                                                                                                                              
COLUMN FAMILIES DESCRIPTION                                                                                                                                        
{NAME => 'f1', BLOOMFILTER => 'ROW',VERSIONS => '1', IN_MEMORY => 'false', KEEP_DELETED_CELLS => 'FALSE',DATA_BLOCK_ENCODING => 'NONE', TTL => 'FOREVER', COMPRESS
ION => 'SNAPPY', MIN_VERSIONS => '0',BLOCKCACHE => 'true', BLOCKSIZE => '65536', REPLICATION_SCOPE =>'0'}                                                         
{NAME => 'f2', BLOOMFILTER => 'ROW',VERSIONS => '1', IN_MEMORY => 'false', KEEP_DELETED_CELLS => 'FALSE',DATA_BLOCK_ENCODING => 'NONE', TTL => 'FOREVER', COMPRESS
ION => 'NONE', MIN_VERSIONS => '0',BLOCKCACHE => 'true', BLOCKSIZE => '65536', REPLICATION_SCOPE =>'0'}                                                          
2 row(s) in 0.0310 seconds
 
hbase(main):009:0> major_compact 'test1'
0 row(s) in 0.1380 seconds
 
hbase(main):010:0> desc 'test1'
Table test1 is ENABLED                                                                                                                                             
test1                                                                                                                                                              
COLUMN FAMILIES DESCRIPTION                                                                                                                                        
{NAME => 'f1', BLOOMFILTER => 'ROW',VERSIONS => '1', IN_MEMORY => 'false', KEEP_DELETED_CELLS => 'FALSE',DATA_BLOCK_ENCODING => 'NONE', TTL => 'FOREVER', COMPRESS
ION => 'SNAPPY', MIN_VERSIONS => '0',BLOCKCACHE => 'true', BLOCKSIZE => '65536', REPLICATION_SCOPE =>'0'}                                                        
{NAME => 'f2', BLOOMFILTER => 'ROW',VERSIONS => '1', IN_MEMORY => 'false', KEEP_DELETED_CELLS => 'FALSE',DATA_BLOCK_ENCODING => 'NONE', TTL => 'FOREVER', COMPRESS
ION => 'NONE', MIN_VERSIONS => '0',BLOCKCACHE => 'true', BLOCKSIZE => '65536', REPLICATION_SCOPE =>'0'}                                                          
2 row(s) in 0.0260 seconds

 

但是没有disable表,也不做major_compact,列族压缩格式也修改成功了(暂时不知道原因)。

hbase(main):001:0> desc 'test'
Table test is ENABLED                                                                                                                                               
test                                                                                                                                                                
COLUMN FAMILIES DESCRIPTION                                                                                                                                         
{NAME => 'fam1', BLOOMFILTER => 'ROW', VERSIONS => '1', IN_MEMORY => 'false', KEEP_DELETED_CELLS => 'FALSE', DATA_BLOCK_ENCODING => 'NONE', TTL => 'FOREVER', COMPRE
SSION => 'NONE', MIN_VERSIONS => '0', BLOCKCACHE => 'true', BLOCKSIZE => '65536', REPLICATION_SCOPE => '0'}                                                         
1 row(s) in 0.3680 seconds

hbase(main):002:0> alter 'test',{NAME=>'fam1',COMPRESSION=>'LZ4'}
Updating all regions with the new schema...
1/1 regions updated.
Done.
0 row(s) in 2.0460 seconds

hbase(main):003:0> desc 'test'
Table test is ENABLED                                                                                                                                               
test                                                                                                                                                               
COLUMN FAMILIES DESCRIPTION                                                                                                                                        
{NAME => 'fam1', BLOOMFILTER =>'ROW', VERSIONS => '1', IN_MEMORY => 'false', KEEP_DELETED_CELLS =>'FALSE', DATA_BLOCK_ENCODING => 'NONE', TTL => 'FOREVER', COMPRE
SSION => 'LZ4', MIN_VERSIONS => '0',BLOCKCACHE => 'true', BLOCKSIZE => '65536', REPLICATION_SCOPE =>'0'}                                                          
1 row(s) in 0.0280 seconds

 

 


网站栏目:hbase的四种压缩格式
标题来源:http://cdkjz.cn/article/jjopds.html
多年建站经验

多一份参考,总有益处

联系快上网,免费获得专属《策划方案》及报价

咨询相关问题或预约面谈,可以通过以下方式与我们联系

大客户专线   成都:13518219792   座机:028-86922220