本文转自http://debugo.com/hbase-shell-cmds/，感谢作者
HBase Shell是HBase的一个命令行工具，我们可以通过它对HBase进行维护操作。我们可以使用sudo -u hbase hbase shell来进入HBase shell。
在HBase shell中，可以使用status, version和whoami分别获得当前服务的状态、版本、登录用户和验证方式。
|
1
2
3
4
5
6
7
|
>
status
3
servers
,
1
dead
,
1.3333
average
load
>
version
0.98.6
-
cdh5
.
3.1
,
rUnknown
,
Tue
Jan
27
16
:
43
:
50
PST
2015
>
whoami
hbase
(
auth
:
SIMPLE
)
groups
:
hbase
|
HBase shell中的帮助命令非常强大，使用help获得全部命令的列表，使用help ‘command_name’获得某一个命令的详细信息。 例如：
|
1
2
3
4
5
6
7
|
>
help
'list'
List
all
tables
in
hbase
.
Optional
regular
expression
parameter
could
be
used
to
filter
the
output
.
Examples
:
hbase
>
list
hbase
>
list
'abc.*'
hbase
>
list
'ns:abc.*'
hbase
>
list
'ns:.*'
|
1. 命名空间
在HBase系统中，命名空间namespace指的是一个HBase表的逻辑分组，同一个命名空间中的表有类似的用途，也用于配额和权限等设置进行安全管控。
HBase默认定义了两个系统内置的预定义命名空间：
• hbase：系统命名空间，用于包含hbase的内部表
• default：所有未指定命名空间的表都自动进入该命名空间
我们可以通过create_namespace命令来建立命名空间
|
1
2
|
>
create
_namespace
'debugo_ns'
0
row
(
s
)
in
2.0910
seconds
|
通过drop_namespace来删除命名空间
|
1
2
|
>
drop
_namespace
'debugo_ns'
0
row
(
s
)
in
1.9540
seconds
|
通过alter_namespac改变表的属性，其格式如下：
alter_namespace 'my_ns', {METHOD => 'set', 'PROPERTY_NAME' => 'PROPERTY_VALUE'}
显示命名空间以及设定的元信息：
|
1
2
3
4
|
>
describe
_namespace
'debugo_ns'
DESCRIPTION
{
NAME
=
>
'debugo_ns'
}
1
row
(
s
)
in
1.9540
seconds
|
显示所有命名空间
|
1
2
3
4
5
6
|
>
list_namespace
NAMESPACE
debugo_ns
default
hbase
3
row
(
s
)
in
0.0910
seconds
|
在HBase下建表需要使用create table_name, column_family1, 这个命令:
|
1
2
3
|
>
create
'user'
,
'info'
0
row
(
s
)
in
0.9030
seconds
=
>
Hbase
::
Table
-
user
|
这个时候这个表是创建在default下面。如果需要在debugo_ns这个命名空间下面建表，则需要使用create namespace:table_name这种方式:
|
1
2
3
4
5
|
>
create
_namespace
'debugo_ns'
0
row
(
s
)
in
2.0910
seconds
create
'debugo_ns:users'
,
'info'
0
row
(
s
)
in
0.4640
seconds
=
>
Hbase
::
Table
-
debugo_ns
:
users
|
List命令可以列出当前HBase实例中的所有表，支持使用正则表达式来匹配。
|
1
2
3
4
|
>
list_namespace
_tables
'debugo_ns'
TABLE
users
1
row
(
s
)
in
0.0400
seconds
|
使用list_namespace_tables也可以直接输出某个命名空间下的所有表
|
1
2
3
4
|
>
list_namespace
_tables
'debugo_ns'
TABLE
users
1
row
(
s
)
in
0.0400
seconds
|
2. DDL语句
首先是建立HBase表，上面我们已经用过create命令了。它后面的第一个参数是表名，然后是一系列列簇的列表。每个列簇中可以独立指定它使用的版本数，数据有效保存时间（TTL），是否开启块缓存等信息。
|
1
|
>
create
't1'
,
{
NAME
=
>
'f1'
,
VERSIONS
=
>
1
,
TTL
=
>
2592000
,
BLOCKCACHE
=
>
true
}
,
'f2'
|
表也可以在创建时指定它预分割(pre-splitting)的region数和split方法。在表初始建立时，HBase只分配给这个表一个region。这就意味着当我们访问这个表数据时，我们只会访问一个region server，这样就不能充分利用集群资源。HBase提供了一个工具来管理表的region数，即org.apache.hadoop.hbase.util.RegionSplitter和HBase shell中create中的split的配置项。例如：
|
1
|
>
create
't2'
,
'f1'
,
{
NUMREGIONS
=
>
3
,
SPLITALGO
=
>
'HexStringSplit'
}
|
我们通过describe 来查看这个表中的元信息：
|
1
2
3
4
|
>
describe
't2'
DESCRIPTION
ENABLED
't2'
,
{
NAME
=
>
'f1'
,
DATA_BLOCK_ENCODING
=
>
'NONE'
,
BLOOMFILTER
=
>
'ROW'
,
REPLIC
true
ATION_SCOPE
=
>
'0'
,
VERSIONS
=
>
'1'
,
COMPRESSION
=
>
'NONE'
,
MIN_VERSIONS
=
>
'0'
,
TTL
=
>
'FOREVER'
,
KEEP_DELETED_CELLS
=
>
'false'
,
BLOCKSIZE
=
>
'65536'
,
IN
_MEMOR
Y
=
>
'false'
,
BLOCKCACHE
=
>
'true'
}
1
row
(
s
)
in
0.0690
seconds
|
通过enable和disable来启用/禁用这个表,相应的可以通过is_enabled和is_disabled来检查表是否被禁用。
|
1
2
3
4
5
6
7
8
9
10
|
>
disable
't2'
0
row
(
s
)
in
1.4410
seconds
>
enable
't2'
0
row
(
s
)
in
0.5940
seconds
>
is
_enabled
't2'
true
0
row
(
s
)
in
0.0400
seconds
hbase
(
main
)
:
042
:
0
>
is
_disabled
't2'
false
0
row
(
s
)
in
0.0490
seconds
|
使用exists来检查表是否存在
|
1
2
3
|
>
exists
't2'
Table
t2
does
exist
0
row
(
s
)
in
0.0590
seconds
|
使用alter来改变表的属性，比如改变列簇的属性, 这涉及将信息更新到所有的region。在过去的版本中，alter操作需要先把table禁用，而在当前版本已经不需要。
|
1
2
3
4
5
6
|
>
alter
't1'
,
{
NAME
=
>
'f1'
,
VERSIONS
=
>
5
}
Updating
all
regions
with
the
new
schema
.
.
.
0
/
1
regions
updated
.
1
/
1
regions
updated
.
Done
.
0
row
(
s
)
in
2.3470
seconds
|
另外一个非常常用的操作是添加和删除列簇：
|
1
2
3
4
5
6
7
|
>
alter
't1'
,
'f3'
Updating
all
regions
with
the
new
schema
.
.
.
0
/
1
regions
updated
.
1
/
1
regions
updated
.
Done
.
0
row
(
s
)
in
2.3130
seconds
>
alter
't1'
,
'delete'
=
>
'f3'
|
或者：
|
1
2
3
4
5
6
|
>
alter
't1'
,
{
NAME
=
>
'f3'
,
METHOD
=
>
'delete'
}
Updating
all
regions
with
the
new
schema
.
.
.
0
/
1
regions
updated
.
1
/
1
regions
updated
.
Done
.
0
row
(
s
)
in
2.2930
seconds
|
删除表需要先将表disable。
|
1
2
3
4
|
>
disable
't1'
0
row
(
s
)
in
1.4310
seconds
>
drop
't1'
0
row
(
s
)
in
0.2440
seconds
|
3. put与get
在HBase shell中，我们可以通过put命令来插入数据。例如我们新创建一个表，它拥有id、address和info三个列簇，并插入一些数据。列簇下的列不需要提前创建，在需要时通过
:
来指定即可。
|
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
|
>
create
'member'
,
'id'
,
'address'
,
'info'
0
row
(
s
)
in
0.4570
seconds
=
>
Hbase
::
Table
–
member
put
'member'
,
'debugo'
,
'id'
,
'11'
put
'member'
,
'debugo'
,
'info:age'
,
'27'
put
'member'
,
'debugo'
,
'info:birthday'
,
'1987-04-04'
put
'member'
,
'debugo'
,
'info:industry'
,
'it'
put
'member'
,
'debugo'
,
'address:city'
,
'beijing'
put
'member'
,
'debugo'
,
'address:country'
,
'china'
put
'member'
,
'Sariel'
,
'id'
,
'21'
put
'member'
,
'Sariel'
,
'info:age'
,
'26'
put
'member'
,
'Sariel'
,
'info:birthday'
,
'1988-05-09 '
put
'member'
,
'Sariel'
,
'info:industry'
,
'it'
put
'member'
,
'Sariel'
,
'address:city'
,
'beijing'
put
'member'
,
'Sariel'
,
'address:country'
,
'china'
put
'member'
,
'Elvis'
,
'id'
,
'22'
put
'member'
,
'Elvis'
,
'info:age'
,
'26'
put
'member'
,
'Elvis'
,
'info:birthday'
,
'1988-09-14 '
put
'member'
,
'Elvis'
,
'info:industry'
,
'it'
put
'member'
,
'Elvis'
,
'address:city'
,
'beijing'
put
'member'
,
'Elvis'
,
'address:country'
,
'china'
|
获取一个id的所有数据
|
1
2
3
4
5
6
7
8
9
|
>
get
'member'
,
'Sariel'
COLUMN
CELL
address
:
city
timestamp
=
1425871035382
,
value
=
beijing
address
:
country
timestamp
=
1425871035424
,
value
=
china
id
:
timestamp
=
1425871035176
,
value
=
21
info
:
age
timestamp
=
1425871035225
,
value
=
26
info
:
birthday
timestamp
=
1425871035296
,
value
=
1988
-
05
-
09
info
:
industry
timestamp
=
1425871035334
,
value
=
it
6
row
(
s
)
in
0.0530
seconds
|
获得一个id，一个列簇（一个列）中的所有数据:
|
1
2
3
4
5
6
7
8
9
10
|
>
get
'member'
,
'Sariel'
,
'info'
COLUMN
CELL
info
:
age
timestamp
=
1425871035225
,
value
=
26
info
:
birthday
timestamp
=
1425871035296
,
value
=
1988
-
05
-
09
info
:
industry
timestamp
=
1425871035334
,
value
=
it
3
row
(
s
)
in
0.0320
seconds
>
get
'member'
,
'Sariel'
,
'info:age'
COLUMN
CELL
info
:
age
timestamp
=
1425871035225
,
value
=
26
1
row
(
s
)
in
0.0270
seconds
|
通过describe ‘member’可以看到，默认情况下列簇只保存1个version。我们先将其修改到2,然后update一些信息。
|
1
2
3
4
5
6
7
8
9
10
11
12
13
|
>
alter
'member'
,
{
NAME
=
>
'info'
,
VERSIONS
=
>
2
}
Updating
all
regions
with
the
new
schema
.
.
.
0
/
1
regions
updated
.
1
/
1
regions
updated
.
Done
.
0
row
(
s
)
in
2.2580
seconds
>
put
'member'
,
'debugo'
,
'info:age'
,
'29'
>
put
'member'
,
'debugo'
,
'info:age'
,
'28'
>
get
'member'
,
'debugo'
,
{
COLUMN
=
>
'info:age'
,
VERSIONS
=
>
2
}
COLUMN
CELL
info
:
age
timestamp
=
1425884510241
,
value
=
28
info
:
age
timestamp
=
1425884510195
,
value
=
29
2
row
(
s
)
in
0.0400
seconds
|
4. 其他DML语句
通过delete命令，我们可以删除id为某个值的‘info:age’字段，接下来的get就无视了
|
1
2
3
4
5
|
>
delete
'member'
,
'debugo'
,
'info:age'
0
row
(
s
)
in
0.0420
seconds
>
get
'member'
,
'debugo'
,
'info:age'
COLUMN
CELL
0
row
(
s
)
in
0.3270
seconds
|
通过deleteall来删除整行
|
1
2
3
4
5
|
>
delete
'member'
,
'debugo'
,
'info:age'
0
row
(
s
)
in
0.0420
seconds
>
get
'member'
,
'debugo'
,
'info:age'
COLUMN
CELL
0
row
(
s
)
in
0.3270
seconds
|
给’Sariel’的’info:age’字段添加，并使用incr实现递增。但需要注意的是，这个value需要是一个数值，如果使用单引号标识的字符串就无法使用incr。在使用Java API开发时，我们可以使用toBytes函数讲数值转换成byte字节。在HBase shell中我们只能通过incr来初始化这个列，
|
1
2
3
4
5
6
7
8
9
10
11
12
|
>
delete
'member'
,
'Sariel'
,
'info:age'
0
row
(
s
)
in
0.0270
seconds
>
incr
'member'
,
'Sariel'
,
'info:age'
,
26
0
row
(
s
)
in
0.0290
seconds
>
incr
'member'
,
'Sariel'
,
'info:age'
0
row
(
s
)
in
0.0290
seconds
>
incr
'member'
,
'Sariel'
,
'info:age'
,
-
1
0
row
(
s
)
in
0.0230
seconds
>
get
'member'
,
'Sariel'
,
'info:age'
COLUMN
CELL
info
:
age
timestamp
=
1425890213341
,
value
=
\
x00
\
x00
\
x00
\
x00
\
x00
\
x00
\
x00
\
x1A
1
row
(
s
)
in
0.0280
seconds
|
十六进制1A是26，通过上面增1再减1后得到的结果。下面通过count统计行数。
|
1
2
3
|
>
count
'member'
2
row
(
s
)
in
0.0750
seconds
=
>
2
|
通过truncate来截断表。hbase是先将掉disable掉，然后drop掉后重建表来实现truncate的功能的。
|
1
2
3
4
5
6
|
hbase
(
main
)
:
010
:
0
>
truncate
'member'
Truncating
'member'
table
(
it
may
take
a
while
)
:
-
Disabling
table
.
.
.
-
Dropping
table
.
.
.
-
Creating
table
.
.
.
0
row
(
s
)
in
2.3260
seconds
|
5. scan和filter
通过scan来对全表进行扫描。我们将之前put的数据恢复。
|
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
|
>
scan
'member'
ROW
COLUMN
+
CELL
Elvis
column
=
address
:
city
,
timestamp
=
1425891057211
,
value
=
beijing
Elvis
column
=
address
:
country
,
timestamp
=
1425891057258
,
val
ue
=
china
Elvis
column
=
id
:
,
timestamp
=
1425891057038
,
value
=
22
Elvis
column
=
info
:
age
,
timestamp
=
1425891057083
,
value
=
26
Elvis
column
=
info
:
birthday
,
timestamp
=
1425891057129
,
value
=
1988
-
09
-
14
Elvis
column
=
info
:
industry
,
timestamp
=
1425891057172
,
value
=
it
Sariel
column
=
address
:
city
,
timestamp
=
1425891056965
,
value
=
beijing
Sariel
column
=
address
:
country
,
timestamp
=
1425891057003
,
val
ue
=
china
Sariel
column
=
id
:
,
timestamp
=
1425891056767
,
value
=
21
Sariel
column
=
info
:
age
,
timestamp
=
1425891056808
,
value
=
26
Sariel
column
=
info
:
birthday
,
timestamp
=
1425891056883
,
value
=
1988
-
05
-
09
Sariel
column
=
info
:
industry
,
timestamp
=
1425891056924
,
value
=
it
debugo
column
=
address
:
city
,
timestamp
=
1425891056642
,
value
=
beijing
debugo
column
=
address
:
country
,
timestamp
=
1425891056726
,
val
ue
=
china
debugo
column
=
id
:
,
timestamp
=
1425891056419
,
value
=
11
debugo
column
=
info
:
age
,
timestamp
=
1425891056499
,
value
=
27
debugo
column
=
info
:
birthday
,
timestamp
=
1425891056547
,
value
=
1987
-
04
-
04
debugo
column
=
info
:
industry
,
timestamp
=
1425891056597
,
value
=
it
3
row
(
s
)
in
0.0660
seconds3
row
(
s
)
in
0.0590
seconds
|
指定扫描其中的某个列：
|
1
|
>
scan
'member'
,
{
COLUMNS
=
>
'info:birthday'
}
|
或者整个列簇：
|
1
2
3
4
5
6
7
8
9
10
11
12
|
>
scan
'member'
,
{
COLUMNS
=
>
'info'
}
ROW
COLUMN
+
CELL
Elvis
column
=
info
:
age
,
timestamp
=
1425891057083
,
value
=
26
Elvis
column
=
info
:
birthday
,
timestamp
=
1425891057129
,
value
=
1988
-
09
-
14
Elvis
column
=
info
:
industry
,
timestamp
=
1425891057172
,
value
=
it
Sariel
column
=
info
:
age
,
timestamp
=
1425891056808
,
value
=
26
Sariel
column
=
info
:
birthday
,
timestamp
=
1425891056883
,
value
=
1988
-
05
-
09
Sariel
column
=
info
:
industry
,
timestamp
=
1425891056924
,
value
=
it
debugo
column
=
info
:
age
,
timestamp
=
1425891056499
,
value
=
27
debugo
column
=
info
:
birthday
,
timestamp
=
1425891056547
,
value
=
1987
-
04
-
04
debugo
column
=
info
:
industry
,
timestamp
=
1425891056597
,
value
=
it
3
row
(
s
)
in
0.0650
seconds
|
除了列（COLUMNS）修饰词外，HBase还支持Limit（限制查询结果行数），STARTROW （ROWKEY起始行。会先根据这个key定位到region，再向后扫描）、STOPROW(结束行)、TIMERANGE（限定时间戳范围）、VERSIONS（版本数）、和FILTER（按条件过滤行）等。比如我们从Sariel这个rowkey开始，找下一个行的最新版本：
|
1
2
3
4
5
6
7
8
9
|
>
scan
'member'
,
{
STARTROW
=
>
'Sariel'
,
LIMIT
=
>
1
,
VERSIONS
=
>
1
}
ROW
COLUMN
+
CELL
Sariel
column
=
address
:
city
,
timestamp
=
1425891056965
,
value
=
beijing
Sariel
column
=
address
:
country
,
timestamp
=
1425891057003
,
value
=
china
Sariel
column
=
id
:
,
timestamp
=
1425891056767
,
value
=
21
Sariel
column
=
info
:
age
,
timestamp
=
1425891056808
,
value
=
26
Sariel
column
=
info
:
birthday
,
timestamp
=
1425891056883
,
value
=
1988
-
05
-
09
Sariel
column
=
info
:
industry
,
timestamp
=
1425891056924
,
value
=
it
1
row
(
s
)
in
0.0410
seconds
|
Filter是一个非常强大的修饰词，可以设定一系列条件来进行过滤。比如我们要限制某个列的值等于26：
|
1
2
3
4
5
|
>
scan
'member'
,
FILTER
=
>
"ValueFilter(=,'binary:26')"
ROW
COLUMN
+
CELL
Elvis
column
=
info
:
age
,
timestamp
=
1425891057083
,
value
=
26
Sariel
column
=
info
:
age
,
timestamp
=
1425891056808
,
value
=
26
2
row
(
s
)
in
0.0620
seconds
|
值包含6这个值：
|
1
2
3
4
|
>
scan
'member'
,
FILTER
=
>
"ValueFilter(=,'substring:6')"
Elvis
column
=
info
:
age
,
timestamp
=
1425891057083
,
value
=
26
Sariel
column
=
info
:
age
,
timestamp
=
1425891056808
,
value
=
26
2
row
(
s
)
in
0.0620
seconds
|
列名中的前缀为birthday的：
|
1
2
3
4
5
6
|
>
scan
'member'
,
FILTER
=
>
"ColumnPrefixFilter('birth') "
ROW
COLUMN
+
CELL
Elvis
column
=
info
:
birthday
,
timestamp
=
1425891057129
,
value
=
1988
-
09
-
14
Sariel
column
=
info
:
birthday
,
timestamp
=
1425891056883
,
value
=
1988
-
05
-
09
debugo
column
=
info
:
birthday
,
timestamp
=
1425891056547
,
value
=
1987
-
04
-
04
3
row
(
s
)
in
0.0450
seconds
|
FILTER中支持多个过滤条件通过括号、AND和OR的条件组合。
|
1
2
3
4
|
>
scan
'member'
,
FILTER
=
>
"ColumnPrefixFilter('birth') AND ValueFilter ValueFilter(=,'substring:1987')"
ROW
COLUMN
+
CELL
Debugo
column
=
info
:
birthday
,
timestamp
=
1425891056547
,
value
=
1987
-
04
-
04
1
row
(
s
)
in
0.0450
seconds
|
同一个rowkey的同一个column有多个version，根据timestamp来区分。而每一个列簇有多个column。而FIRSTKEYONLY仅取出每个列簇的第一个column的第一个版本。而KEYONLY则是对于每一个column只去取出key，把VALUE的信息丢弃,一般和其他filter结合使用。例如：
|
1
2
3
4
5
6
7
8
9
10
11
12
|
>
scan
'member'
,
FILTER
=
>
"FirstKeyOnlyFilter()"
ROW
COLUMN
+
CELL
Elvis
column
=
address
:
city
,
timestamp
=
1425891057211
,
value
=
beijing
Sariel
column
=
address
:
city
,
timestamp
=
1425891056965
,
value
=
beijing
debugo
column
=
address
:
city
,
timestamp
=
1425891056642
,
value
=
beijing
3
row
(
s
)
in
0.0230
seconds
>
scan
'member'
,
FILTER
=
>
"KeyOnlyFilter()"
hbase
(
main
)
:
055
:
0
>
scan
'member'
,
FILTER
=
>
"KeyOnlyFilter()"
ROW
COLUMN
+
CELL
Elvis
column
=
address
:
city
,
timestamp
=
1425891057211
,
value
=
Elvis
column
=
id
:
,
timestamp
=
1425891057038
,
value
=
……
|
PrefixFilter是对Rowkey的前缀进行判断,这是一个非常常用的功能。
|
1
2
3
4
5
|
>
scan
'member'
,
FILTER
=
>
"PrefixFilter('E')"
ROW
COLUMN
+
CELL
Elvis
column
=
address
:
city
,
timestamp
=
1425891057211
,
value
=
beijing
……
1
row
(
s
)
in
0.0460
seconds
|
^^
6万+
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
