HBase过滤器种类很多,我们选择8种常用的过滤器进行介绍。为了获得更好的示例效果,先利用HBase Shell新建students表格,并往表格中进行写入多行数据。
一、数据准备工作
(1)在默认命名空间中新建表格students,设置列族info、score。
hbase:002:0> create 'students','info','score'
2024-03-26 00:22:15,810 INFO [main] client.HBaseAdmin (HBaseAdmin.java:postOperationResult(3591)) - Operation: CREATE, Table Name: default:students, procId: 290 completed
Created table students
Took 3.1425 seconds
=> Hbase::Table - students
(2)往students表格中写入5行数据,并用scan 'students'命令查看写入结果。
hbase:005:0> put 'students','s001','info:name','Jack'
Took 30.6978 seconds
hbase:017:0> put 'students','s001','info:age','18'
Took 0.0419 seconds
hbase:019:0> put 'students','s001','score:English','95'
Took 0.0472 seconds
hbase:021:0> put 'students','s002','info:name','Tom'
Took 0.0255 seconds
hbase:022:0> put 'students','s002','info:age','20'
Took 0.0160 seconds
hbase:023:0> put 'students','s002','score:Chinese','85'
Took 0.0296 seconds
hbase:024:0> put 'students','s002','score:Math','90'
Took 0.0155 seconds
hbase:025:0> put 'students','s003','info:name','Mike'
Took 0.0188 seconds
hbase:026:0> put 'students','s003','info:age','19'
Took 0.0183 seconds
hbase:027:0> put 'students','s003','score:Chinese','90'
Took 0.0178 seconds
hbase:028:0> put 'students','s003','score:Math','95'
Took 0.0445 seconds
hbase:029:0> put 'students','s004','info:name','Lucy'
Took 0.0104 seconds
hbase:030:0> put 'students','s004','score:English','100'
Took 0.0170 seconds
hbase:031:0> put 'students','s005','info:name','Lily'
Took 0.0249 seconds
hbase:032:0> put 'students','s005','score:Chinese','99'
Took 0.0228 seconds
hbase:033:0> scan 'students'
ROW COLUMN+CELL
s001 column=info:age, timestamp=2024-03-26T00:25:17.982, value=18
s001 column=info:name, timestamp=2024-03-26T00:24:39.510, value=Jack
s001 column=score:English, timestamp=2024-03-26T00:25:52.207, value=95
s002 column=info:age, timestamp=2024-03-26T00:26:46.922, value=20
s002 column=info:name, timestamp=2024-03-26T00:26:26.924, value=Tom
s002 column=score:Chinese, timestamp=2024-03-26T00:27:13.181, value=85
s002 column=score:Math, timestamp=2024-03-26T00:27:28.787, value=90
s003 column=info:age, timestamp=2024-03-26T00:28:08.402, value=19
s003 column=info:name, timestamp=2024-03-26T00:27:48.629, value=Mike
s003 column=score:Chinese, timestamp=2024-03-26T00:28:46.714, value=90
s003 column=score:Math, timestamp=2024-03-26T00:29:01.881, value=95
s004 column=info:name, timestamp=2024-03-26T00:29:19.868, value=Lucy
s004 column=score:English, timestamp=2024-03-26T00:29:44.831, value=100
s005 column=info:name, timestamp=2024-03-26T00:30:04.231, value=Lily
s005 column=score:Chinese, timestamp=2024-03-26T00:30:25.477, value=99
5 row(s)
Took 0.3369 seconds
二、过滤器的使用介绍
1.ValueFilter过滤器
根据数据列单元格的值进行过滤。值过滤器的比较方式有二进制位比较(binary)和子字符串匹配比较(substring)。
(1)按二进制位进行值比较
使用get命令,查询students表格中,行键为s001,单元格值为Jack的数据结果。
#ValueFilter(=,'binary:Jack')是值过滤器,比较方式是binary二进制
hbase:034:0> get 'students','s001',{FILTER=>"ValueFilter(=,'binary:Jack')"}
COLUMN CELL
info:name timestamp=2024-03-26T00:24:39.510, value=Jack
1 row(s)
Took 0.6506 seconds
使用scan命令,扫描出students表格中,单元格值为90的数据结果。
#查询结果是多条,需要用scan命令全表扫描,不能使用get命令
hbase:036:0> scan 'students',{FILTER=>"ValueFilter(=,'binary:90')"}
ROW COLUMN+CELL
s002 column=score:Math, timestamp=2024-03-26T00:27:28.787, value=90
s003 column=score:Chinese, timestamp=2024-03-26T00:28:46.714, value=90
2 row(s)
Took 0.2162 seconds
(2)按子字符串匹配比较
使用get命令,查询students表格中,行键为s001,单元格值包含子字符串ac的数据结果。
hbase:037:0> get 'students','s001',{FILTER=>"ValueFilter(=,'substring:ac')"}
COLUMN CELL
info:name timestamp=2024-03-26T00:24:39.510, value=Jack
1 row(s)
Took 0.1578 seconds
使用scan命令,扫描出表格students中单元格值包含子字符串0的数据结果。
#查询结果是多条,需要用scan命令全表扫描,不能使用get命令
hbase:038:0> scan 'students',{FILTER=>"ValueFilter(=,'substring:0')"}
ROW COLUMN+CELL
s002 column=info:age, timestamp=2024-03-26T00:26:46.922, value=20
s002 column=score:Math, timestamp=2024-03-26T00:27:28.787, value=90
s003 column=score:Chinese, timestamp=2024-03-26T00:28:46.714, value=90
s004 column=score:English, timestamp=2024-03-26T00:29:44.831, value=100
3 row(s)
Took 0.0868 seconds
2.QualifierFilter过滤器
列限定符过滤器QualifierFilter是只根据数据列的列限定符进行过滤,并不关注列族名称。列限定符过滤器的常用比较方式为二进制位(binary)比较。
使用get命令,查询students表格中,行键为s001,列限定符为name的数据结果。
hbase:039:0> get 'students','s001',{FILTER=>"QualifierFilter(=,'binary:name')"}
COLUMN CELL
info:name timestamp=2024-03-26T00:24:39.510, value=Jack
1 row(s)
Took 0.3310 seconds
使用scan命令,扫描students表格中,列限定符为name的数据结果。
hbase:041:0> scan 'students',{FILTER=>"QualifierFilter(=,'binary:name')"}
ROW COLUMN+CELL
s001 column=info:name, timestamp=2024-03-26T00:24:39.510, value=Jack
s002 column=info:name, timestamp=2024-03-26T00:26:26.924, value=Tom
s003 column=info:name, timestamp=2024-03-26T00:27:48.629, value=Mike
s004 column=info:name, timestamp=2024-03-26T00:29:19.868, value=Lucy
s005 column=info:name, timestamp=2024-03-26T00:30:04.231, value=Lily
5 row(s)
Took 0.0845 seconds
3.ColumnPrefixFilter过滤器
列前缀符过滤器ColumnPrefixFilter是根据数据列的列限定符的前缀进行过滤。前缀过滤必须从第一个字符开始匹配,而子字符串过滤可以从任何位置开始进行子串匹配。前缀过滤器严格区分字母大小写。
使用get命令,查询出students表格中,行键为s002,列限定符的前缀字符串为Chi的数据结果。
hbase:042:0> get 'students','s002',{FILTER=>"ColumnPrefixFilter('Chi')"}
COLUMN CELL
score:Chinese timestamp=2024-03-26T00:27:13.181, value=85
1 row(s)
Took 0.1693 seconds
使用scan命令,扫描students表格,列限定符的前缀字符串为Chi的数据结果。
hbase:044:0> scan 'students',{FILTER=>"ColumnPrefixFilter('Chi')"}
ROW COLUMN+CELL
s002 column=score:Chinese, timestamp=2024-03-26T00:27:13.181, value=85
s003 column=score:Chinese, timestamp=2024-03-26T00:28:46.714, value=90
s005 column=score:Chinese, timestamp=2024-03-26T00:30:25.477, value=99
3 row(s)
Took 0.0397 seconds
4.RowFilter过滤器
行键过滤器RowFilter是根据行键对数据列进行过滤。
注意:一般不在get命令中使用行键过滤器,get命令必须指定唯一确定完整的行键,没有必要再对行键进行过滤。
(1)按二进制位比较。
使用scan命令,扫描students表格,筛选出行键值为s001的所有数据结果。
hbase:045:0> scan 'students',{FILTER=>"RowFilter(=,'binary:s001')"}
ROW COLUMN+CELL
s001 column=info:age, timestamp=2024-03-26T00:25:17.982, value=18
s001 column=info:name, timestamp=2024-03-26T00:24:39.510, value=Jack
s001 column=score:English, timestamp=2024-03-26T00:25:52.207, value=95
1 row(s)
Took 0.1297 seconds
(2)按子字符串匹配比较。
使用scan命令,扫描students表格,筛选出行键值包含子字符串01的所有数据结果。
hbase:046:0> scan 'students',{FILTER=>"RowFilter(=,'substring:01')"}
ROW COLUMN+CELL
s001 column=info:age, timestamp=2024-03-26T00:25:17.982, value=18
s001 column=info:name, timestamp=2024-03-26T00:24:39.510, value=Jack
s001 column=score:English, timestamp=2024-03-26T00:25:52.207, value=95
1 row(s)
Took 0.3426 seconds
5.PrefixFilter过滤器
行键前缀过滤器PrefixFilter是根据行键的前缀进行过滤。前缀过滤必须从行键的第一个字符开始匹配,严格区分字母大小写。
使用scan命令,扫描students表格,筛选出行键值以s00为前缀开头的数据结果。
hbase:047:0> scan 'students',{FILTER=>"PrefixFilter('s00')"}
ROW COLUMN+CELL
s001 column=info:age, timestamp=2024-03-26T00:25:17.982, value=18
s001 column=info:name, timestamp=2024-03-26T00:24:39.510, value=Jack
s001 column=score:English, timestamp=2024-03-26T00:25:52.207, value=95
s002 column=info:age, timestamp=2024-03-26T00:26:46.922, value=20
s002 column=info:name, timestamp=2024-03-26T00:26:26.924, value=Tom
s002 column=score:Chinese, timestamp=2024-03-26T00:27:13.181, value=85
s002 column=score:Math, timestamp=2024-03-26T00:27:28.787, value=90
s003 column=info:age, timestamp=2024-03-26T00:28:08.402, value=19
s003 column=info:name, timestamp=2024-03-26T00:27:48.629, value=Mike
s003 column=score:Chinese, timestamp=2024-03-26T00:28:46.714, value=90
s003 column=score:Math, timestamp=2024-03-26T00:29:01.881, value=95
s004 column=info:name, timestamp=2024-03-26T00:29:19.868, value=Lucy
s004 column=score:English, timestamp=2024-03-26T00:29:44.831, value=100
s005 column=info:name, timestamp=2024-03-26T00:30:04.231, value=Lily
s005 column=score:Chinese, timestamp=2024-03-26T00:30:25.477, value=99
5 row(s)
Took 0.4404 seconds
6.FamilyFilter过滤器
列族过滤器FamilyFilter是根据列族名称进行过滤。列族过滤器的比较方式有二进制位比较(binary)、子字符串匹配比较(substring)等。
(1)按二进制位比较。
使用scan命令,扫描表格students,筛选出列族名称值为info的数据结果。
hbase:005:0> scan 'students',FILTER=>"FamilyFilter(=,'binary:info')"
ROW COLUMN+CELL
s001 column=info:age, timestamp=2024-03-26T00:25:17.982, value=18
s001 column=info:name, timestamp=2024-03-26T00:24:39.510, value=Jack
s002 column=info:age, timestamp=2024-03-26T00:26:46.922, value=20
s002 column=info:name, timestamp=2024-03-26T00:26:26.924, value=Tom
s003 column=info:age, timestamp=2024-03-26T00:28:08.402, value=19
s003 column=info:name, timestamp=2024-03-26T00:27:48.629, value=Mike
s004 column=info:name, timestamp=2024-03-26T00:29:19.868, value=Lucy
s005 column=info:name, timestamp=2024-03-26T00:30:04.231, value=Lily
5 row(s)
Took 0.0399 seconds
(2)按子字符串匹配比较。
使用scan命令,扫描表格students,筛选出列族名称包含子字符串s的数据结果。
hbase:008:0> scan 'students',FILTER=>"FamilyFilter(=,'substring:s')"
ROW COLUMN+CELL
s001 column=score:English, timestamp=2024-03-26T00:25:52.207, value=95
s002 column=score:Chinese, timestamp=2024-03-26T00:27:13.181, value=85
s002 column=score:Math, timestamp=2024-03-26T00:27:28.787, value=90
s003 column=score:Chinese, timestamp=2024-03-26T00:28:46.714, value=90
s003 column=score:Math, timestamp=2024-03-26T00:29:01.881, value=95
s004 column=score:English, timestamp=2024-03-26T00:29:44.831, value=100
s005 column=score:Chinese, timestamp=2024-03-26T00:30:25.477, value=99
5 row(s)
Took 0.0915 seconds
7.SingleColumnValueFilter过滤器
单列值过滤器SingleColumnValueFilters是根据指定列族和列限定符的单个数据列的单元格值进行过滤,类似SQL中的”select列名from表名where列名=值”语句。
(1)按二进制位比较。
使用scan命令,扫描表格students,筛选出列族info,列限定符age的单元格值为19的数据列。
hbase:006:0> scan 'students',{COLUMN=>'info:age',FILTER=>"SingleColumnValueFilter('info','age',=,'binary:19')"}
ROW COLUMN+CELL
s003 column=info:age, timestamp=2024-03-26T00:28:08.402, value=19
1 row(s)
Took 0.4166 seconds
(2)按子字符串匹配比较。
使用scan命令,扫描表格students,筛选出列族info,列限定符name的值包括子字符串y的数据。
hbase:008:0> scan 'students',{COLUMN=>'info:name',FILTER=>"SingleColumnValueFilter('info','name',=,'substring:y')"}
ROW COLUMN+CELL
s004 column=info:name, timestamp=2024-03-26T00:29:19.868, value=Lucy
s005 column=info:name, timestamp=2024-03-26T00:30:04.231, value=Lily
2 row(s)
Took 0.0658 seconds