接上一篇，HBase与传统数据库一个很大的不同之处：HBase可以保存多个版本的值，而不仅仅是保存最新的值。版本之间，通过timestamp属性来区分。
我们来具体看一下，首先新增一行visit100，我们多次设置列personinfo:name，每次设置后都查询一次最新的值：
hbase(main):011:0> put 'patientvisit','visit100','personinfo:name','A' 0 row(s) in 0.0890 seconds hbase(main):012:0> get 'patientvisit','visit100','personinfo:name' COLUMN CELL personinfo:name timestamp=1454342758182, value=A 1 row(s) in 0.0310 seconds hbase(main):013:0> put 'patientvisit','visit100','personinfo:name','B' 0 row(s) in 0.0100 seconds hbase(main):014:0> get 'patientvisit','visit100','personinfo:name' COLUMN CELL personinfo:name timestamp=1454342796166, value=B 1 row(s) in 0.0230 seconds hbase(main):015:0> put 'patientvisit','visit100','personinfo:name','B' 0 row(s) in 0.0140 seconds hbase(main):016:0> get 'patientvisit','visit100','personinfo:name' COLUMN CELL personinfo:name timestamp=1454342802829, value=B 1 row(s) in 0.0120 seconds
我们可以看到，每次设置该列的数值timestamp都有相应的变化：
| timestamp | value |
| 1454342758182 | A |
| 1454342796166 | B |
| 1454342802829 | B |
查询的时候，默认返回最新一次的值，当然我们也可以查询对应版本的数据：
注意：如果没有该timestamp，则不返回数据
hbase(main):004:0> get 'patientvisit','visit100',COLUMN=>'personinfo:name',TIMESTAMP=>1454342758182 COLUMN CELL personinfo:name timestamp=1454342758182, value=A 1 row(s) in 0.0110 seconds hbase(main):005:0> get 'patientvisit','visit100',COLUMN=>'personinfo:name',TIMESTAMP=>1454342758183 COLUMN CELL 0 row(s) in 0.0130 seconds
查询的时候，我们也可以通过TIMERANGE来查询该时间段内的最新值：
hbase(main):002:0> get 'patientvisit','visit100',COLUMN=>'personinfo:name',TIMERANGE=>[1454342758182,1454342796166] COLUMN CELL personinfo:name timestamp=1454342758182, value=A 1 row(s) in 0.0200 seconds hbase(main):003:0> get 'patientvisit','visit100',COLUMN=>'personinfo:name',TIMERANGE=>[1454342758182,1454342796167] COLUMN CELL personinfo:name timestamp=1454342796166, value=B 1 row(s) in 0.0240 seconds
然后再说一下版本问题，为了说明这个问题，我们要先修改一下表的SCHEMA：
hbase(main):013:0> describe 'patientvisit'
Table patientvisit is ENABLED
patientvisit
COLUMN FAMILIES DESCRIPTION
{NAME => 'personinfo', DATA_BLOCK_ENCODING => 'NONE', BLOOMFILTER => 'ROW', REPLICATION_SCOPE => '0', VERSIONS => '1', COMPRESSION => 'NONE', MIN_VERSIONS =>
'0', TTL => 'FOREVER', KEEP_DELETED_CELLS => 'FALSE', BLOCKSIZE => '65536', IN_MEMORY => 'false', BLOCKCACHE => 'true'}
{NAME => 'personinfoex', DATA_BLOCK_ENCODING => 'NONE', BLOOMFILTER => 'ROW', REPLICATION_SCOPE => '0', VERSIONS => '1', COMPRESSION => 'NONE', MIN_VERSIONS
=> '0', TTL => 'FOREVER', KEEP_DELETED_CELLS => 'FALSE', BLOCKSIZE => '65536', IN_MEMORY => 'false', BLOCKCACHE => 'true'}
{NAME => 'visitinfo', DATA_BLOCK_ENCODING => 'NONE', BLOOMFILTER => 'ROW', REPLICATION_SCOPE => '0', VERSIONS => '1', COMPRESSION => 'NONE', MIN_VERSIONS =>
'0', TTL => 'FOREVER', KEEP_DELETED_CELLS => 'FALSE', BLOCKSIZE => '65536', IN_MEMORY => 'false', BLOCKCACHE => 'true'}
{NAME => 'visitinfoex', DATA_BLOCK_ENCODING => 'NONE', BLOOMFILTER => 'ROW', REPLICATION_SCOPE => '0', VERSIONS => '1', COMPRESSION => 'NONE', MIN_VERSIONS =
> '0', TTL => 'FOREVER', KEEP_DELETED_CELLS => 'FALSE', BLOCKSIZE => '65536', IN_MEMORY => 'false', BLOCKCACHE => 'true'}
4 row(s) in 0.0510 seconds
hbase(main):020:0> alter 'patientvisit',{NAME=>'personinfoex',VERSIONS=>5}
Updating all regions with the new schema...
0/1 regions updated.
1/1 regions updated.
Done.
0 row(s) in 3.2850 seconds
然后，对于personinfoex:money进行重复赋值
hbase(main):022:0> put 'patientvisit','visit100','personinfoex:money','100' 0 row(s) in 0.2510 seconds hbase(main):023:0> put 'patientvisit','visit100','personinfoex:money','1000' 0 row(s) in 0.0140 seconds hbase(main):024:0> put 'patientvisit','visit100','personinfoex:money','10000' 0 row(s) in 0.0090 seconds hbase(main):025:0> put 'patientvisit','visit100','personinfoex:money','100000' 0 row(s) in 0.0110 seconds hbase(main):026:0> put 'patientvisit','visit100','personinfoex:money','1000000' 0 row(s) in 0.0110 seconds hbase(main):027:0> put 'patientvisit','visit100','personinfoex:money','10000000' 0 row(s) in 0.0160 seconds
然后，get的时候，传入VERSIONS参数：
hbase(main):029:0> get 'patientvisit','visit100',COLUMN=>'personinfoex:money' COLUMN CELL personinfoex:money timestamp=1454466774726, value=10000000 1 row(s) in 0.0690 seconds hbase(main):031:0> get 'patientvisit','visit100',COLUMN=>'personinfoex:money',VERSIONS=>1 COLUMN CELL personinfoex:money timestamp=1454466774726, value=10000000 1 row(s) in 0.0160 seconds hbase(main):032:0> get 'patientvisit','visit100',COLUMN=>'personinfoex:money',VERSIONS=>2 COLUMN CELL personinfoex:money timestamp=1454466774726, value=10000000 personinfoex:money timestamp=1454466769620, value=1000000 2 row(s) in 0.0100 seconds hbase(main):030:0> get 'patientvisit','visit100',COLUMN=>'personinfoex:money',VERSIONS=>3 COLUMN CELL personinfoex:money timestamp=1454466774726, value=10000000 personinfoex:money timestamp=1454466769620, value=1000000 personinfoex:money timestamp=1454466766192, value=100000 3 row(s) in 0.0270 seconds hbase(main):033:0> get 'patientvisit','visit100',COLUMN=>'personinfoex:money',VERSIONS=>4 COLUMN CELL personinfoex:money timestamp=1454466774726, value=10000000 personinfoex:money timestamp=1454466769620, value=1000000 personinfoex:money timestamp=1454466766192, value=100000 personinfoex:money timestamp=1454466763838, value=10000 4 row(s) in 0.0410 seconds
恩，那再总结一下，HBase就是一个有时间戳和版本管理的三层KV数据库。