list 查看表 带有正则写法:
hbase(main):014:0> list 'zm.*'
TABLE
zmtest1
1 row(s) in 0.0060 seconds
=> ["zmtest1"]
如下是案例中总结出来的命令:
rowkey: PrefixFilter
列名:ColumnPrefixFilter
列值:ValueFilter (精确/模糊)
ValueFilter(>,'binary:\x00\x00\x00\x00\x00\x00\x00\x0F')
ValueFilter(=,'binary:sku188')
ValueFilter(=,'substring:1')
可以结合limit使用 LIMIT => 10
scan 'zmtest1', LIMIT => 10, FILTER=>"( ColumnPrefixFilter('app') AND ValueFilter(>,'binary:\x00\x00\x00\x00\x00\x00\x00\x0F') )"
scan 'zmtest1', LIMIT => 10 查看10条记录
FirstKeyOnlyFilter: 只获取hbase表第一个列和对应value值
RowFilter: 只获取rowkey
hbase 列值为binary下映射到hive:
CREATE EXTERNAL TABLE user_app_cookie_list ( username STRING, app1_cookie_id BIGINT, app2_cookie_id BIGINT )
STORED BY 'org.apache.hadoop.hive.hbase.HBaseStorageHandler'
WITH SERDEPROPERTIES ("hbase.columns.mapping" = ":key, lf:app1#b, lf:app2#b")
TBLPROPERTIES("hbase.table.name" = "zmtest1");
因为下面案例 hbase shell put的值都是二进制,因此在创建hive外表时,需要指定 #b,否则
关联后查询的hive列值是null.
而在案例 hbase+hive结合使用 中,因为shell put的数据都是字符串因此创建外表
不需要指定格式,曾实验用#b #s创建hive外表,查询结果都是正确的,
估计默认就是#s格式, #b只在hbase shell中显示插入二进制数据时才用到
-- 增加列族名
alter 'zmtest1', NAME => 'cf'
-- 删除列族
alter 'zmtest1', NAME => 'cf', METHOD => 'delete'
或者
alter 'zmtest1', 'delete' => 'cf'
=============如下是案例============================
create 'zmtest1', 'lf', 'sf'
-- lf: column family of LONG values (binary value)
-- sf: column family of STRING values
-- 一个用户(userX),在什么时间(tsX),作为rowkey
-- 对什么产品(value:skuXXX),做了什么操作作为列名,比如,c1: click from homepage; c2: click from ad; s1: search from homepage; b1: buy
put 'zmtest1', 'user1|ts1', 'sf:c1', 'sku1' -- 向表zmtest1中,rowkey为user1|ts1的记录,添加列族为sf下列名为c1,列值为sku1
put 'zmtest1', 'user1|ts2', 'sf:c1', 'sku188'
put 'zmtest1', 'user1|ts3', 'sf:s1', 'sku123'
put 'zmtest1', 'user2|ts4', 'sf:c1', 'sku2'
put 'zmtest1', 'user2|ts5', 'sf:c2', 'sku288'
put 'zmtest1', 'user2|ts6', 'sf:s1', 'sku222'
====================================hbase shell filter=======================================================
scan 'zmtest1', FILTER=>"ValueFilter(=,'binary:sku188')" -- 过滤列值value, binary表示列值二进制下的精确查找
hbase(main):012:0> scan 'zmtest1',FILTER=>"ValueFilter(=,'binary:sku188')"
ROW COLUMN+CELL
user1|ts2 column=sf:c1, timestamp=1468893517729, value=sku188
scan 'zmtest1', FILTER=>"ValueFilter(=,'substring:1')"
hbase(main):013:0> scan 'zmtest1',FILTER=>"ValueFilter(=,'substring:1')" substring表示列值的截取串查找
ROW COLUMN+CELL
user1|ts1 column=sf:c1, timestamp=1468893512967, value=sku1
user1|ts2 column=sf:c1, timestamp=1468893517729, value=sku188
user1|ts3 column=sf:s1, timestamp=1468893522589, value=sku123
scan 'zmtest1', FILTER=>"ValueFilter(=,'substring:88')"
scan 'zmtest1', FILTER=>"ColumnPrefixFilter('c2') AND ValueFilter(=,'substring:88')" -- ColumnPrefixFilter表示列名的substring包含,列名和列值的组合过滤
scan 'zmtest1', FILTER=>"ColumnPrefixFilter('s') AND ( ValueFilter(=,'substring:123') OR ValueFilter(=,'substring:222') )"
=============================================================================================================
c, 一个rowkey可以有多个版本,以timestamp来区别,而且同一个key 会有多个列名,column, 只拿出key里面第一个column的第一个version版本的值,
KeyOnlyFilter(): 只拿出key(keyonly), 去掉value的获取 这样在做count的操作时,就提高效率。
scan 'zmtest1', FILTER=>"FirstKeyOnlyFilter() AND ValueFilter(=,'binary:sku188') AND KeyOnlyFilter()"
--结果
ROW COLUMN+CELL
user1|ts2 column=sf:c1, timestamp=1468893517729, value=
scan 'zmtest1', FILTER=>"FirstKeyOnlyFilter() AND ValueFilter(=,'binary:sku188')"
--结果
ROW COLUMN+CELL
user1|ts2 column=sf:c1, timestamp=1468893517729, value=sku188
-- rowkey过滤PrefixFilter
scan 'zmtest1', FILTER => "PrefixFilter('user1')"
--结果:
ROW COLUMN+CELL
user1|ts1 column=sf:c1, timestamp=1468893512967, value=sku1
user1|ts2 column=sf:c1, timestamp=1468893517729, value=sku188
user1|ts3 column=sf:s1, timestamp=1468893522589, value=sku123
-- STARTROW表示从这一行开始,包含这一行
-- STOPROW 表示到这行结束,不包含这一样
-- 其中,startrow stoprow的数值不一定就是精确值,可以是rowkey里存在的某一个子串
scan 'zmtest1', {STARTROW=>'user1|ts2', FILTER => "PrefixFilter ('user1')"}
--结果
ROW COLUMN+CELL
user1|ts2 column=sf:c1, timestamp=1468893517729, value=sku188
user1|ts3 column=sf:s1, timestamp=1468893522589, value=sku123
scan 'zmtest1', {STARTROW=>'user1|ts2', STOPROW=>'user2'}
--结果
ROW COLUMN+CELL
user1|ts2 column=sf:c1, timestamp=1468893517729, value=sku188
user1|ts3 column=sf:s1, timestamp=1468893522589, value=sku123
scan 'zmtest1', {STARTROW=>'user1', STOPROW=>'user2'} -- user2不一定就是一个精确的rowkey
--结果
ROW COLUMN+CELL
user1|ts1 column=sf:c1, timestamp=1468893512967, value=sku1
user1|ts2 column=sf:c1, timestamp=1468893517729, value=sku188
user1|ts3 column=sf:s1, timestamp=1468893522589, value=sku123
==========================上面的filter是直接可以调用的,下面的filter是需要引用后才能使用===============================
--
import org.apache.hadoop.hbase.filter.CompareFilter
import org.apache.hadoop.hbase.filter.SubstringComparator
import org.apache.hadoop.hbase.filter.RowFilter
-- rowkey里面包含ts3
scan 'zmtest1', {FILTER => RowFilter.new(CompareFilter::CompareOp.valueOf('EQUAL'), SubstringComparator.new('ts3'))}
eg:
hbase(main):024:0> import org.apache.hadoop.hbase.filter.CompareFilter
=> Java::OrgApacheHadoopHbaseFilter::CompareFilter
hbase(main):025:0> import org.apache.hadoop.hbase.filter.SubstringComparator
=> Java::OrgApacheHadoopHbaseFilter::SubstringComparator
hbase(main):026:0> import org.apache.hadoop.hbase.filter.RowFilter
=> Java::OrgApacheHadoopHbaseFilter::RowFilter
hbase(main):027:0> scan 'zmtest1', {FILTER => RowFilter.new(CompareFilter::CompareOp.valueOf('EQUAL'), SubstringComparator.new('ts3'))}
ROW COLUMN+CELL
user1|ts3 column=sf:s1, timestamp=1468893522589, value=sku123
-- 引用正则
import org.apache.hadoop.hbase.filter.RegexStringComparator
put 'zmtest1', 'user2|err', 'sf:s1', 'sku999'
scan 'zmtest1', {FILTER => RowFilter.new(CompareFilter::CompareOp.valueOf('EQUAL'),RegexStringComparator.new('^user\d+\|ts\d+$'))}
--结果
ROW COLUMN+CELL
user1|ts1 column=sf:c1, timestamp=1468893512967, value=sku1
user1|ts2 column=sf:c1, timestamp=1468893517729, value=sku188
user1|ts3 column=sf:s1, timestamp=1468893522589, value=sku123
user2|ts4 column=sf:c1, timestamp=1468893527021, value=sku2
user2|ts5 column=sf:c2, timestamp=1468893527046, value=sku288
user2|ts6 column=sf:s1, timestamp=1468893528197, value=sku222
import org.apache.hadoop.hbase.filter.CompareFilter
import org.apache.hadoop.hbase.filter.SingleColumnValueFilter
import org.apache.hadoop.hbase.filter.SubstringComparator
import org.apache.hadoop.hbase.util.Bytes
scan 't1', { COLUMNS => 'family:qualifier', FILTER =>
SingleColumnValueFilter.new
(Bytes.toBytes('family'),
Bytes.toBytes('qualifier'),
CompareFilter::CompareOp.valueOf('EQUAL'),
SubstringComparator.new('somevalue'))
}
put 'zmtest1', 'user1|ts9', 'sf:b1', 'sku1'
scan 'zmtest1', FILTER=>"ColumnPrefixFilter('b1') AND ValueFilter(=,'binary:sku1')"
scan 'zmtest1', {COLUMNS => 'sf:b1', FILTER => SingleColumnValueFilter.new(Bytes.toBytes('sf'), Bytes.toBytes('b1'), CompareFilter::CompareOp.valueOf('EQUAL'), Bytes.toBytes('sku1'))}
-- binary value --
org.apache.hadoop.hbase.util.Bytes.toString("Hello HBase".to_java_bytes)
org.apache.hadoop.hbase.util.Bytes.toString("\x48\x65\x6c\x6c\x6f\x20\x48\x42\x61\x73\x65".to_java_bytes)
-- 用户userX,作为rowkey,他的各种设备(brwoser, app, pc)作为列名,所对应的cookie_id作为value (长整型变量)
put 'zmtest1', 'user1', 'lf:browser1', "\x00\x00\x00\x00\x00\x00\x00\x02"
put 'zmtest1', 'user1', 'lf:app1', "\x00\x00\x00\x00\x00\x00\x00\x0F"
put 'zmtest1', 'user1', 'lf:app2', "\x00\x00\x00\x00\x00\x00\x00\x10"
put 'zmtest1', 'user2', 'lf:app1', "\x00\x00\x00\x00\x00\x00\x00\x11"
put 'zmtest1', 'user2', 'lf:pc1', "\x00\x00\x00\x00\x00\x00\x00\x12"
-- ValueFilter后可以跟着 = > < 等符号
scan 'zmtest1', {STOPROW=>'user2', FILTER=>"( ColumnPrefixFilter('app') AND ValueFilter(>,'binary:\x00\x00\x00\x00\x00\x00\x00\x0F') )"}
scan 'zmtest1', LIMIT => 10, FILTER=>"( ColumnPrefixFilter('app') AND ValueFilter(>,'binary:\x00\x00\x00\x00\x00\x00\x00\x0F') )"
--结果
ROW COLUMN+CELL
user1 column=lf:app2, timestamp=1468906199761, value=\x00\x00\x00\x00\x00\x00\x00\x10
user2 column=lf:app1, timestamp=1468906199806, value=\x00\x00\x00\x00\x00\x00\x00\x11
-- 增加列族名
alter 'zmtest1', NAME => 'cf'
-- 删除列族
alter 'zmtest1', NAME => 'cf', METHOD => 'delete'
或者
alter 'zmtest1', 'delete' => 'cf'
-- desc 'zmtest1'
-- hive hbase mapping -- 其中,:key可以忽略掉
CREATE EXTERNAL TABLE user_app_cookie_list ( username STRING, app1_cookie_id BIGINT, app2_cookie_id BIGINT )
STORED BY 'org.apache.hadoop.hive.hbase.HBaseStorageHandler'
WITH SERDEPROPERTIES ("hbase.columns.mapping" = ":key, lf:app1#b, lf:app2#b")
TBLPROPERTIES("hbase.table.name" = "zmtest1");
select * from user_app_cookie_list;
--
#b 表示binary
#s 表示字符串
是base.columns.mapping的关键字
--
hbase符合上述条件的数据:
user1 column=lf:app1, timestamp=1468906199730, value=\x00\x00\x00\x00\x00\x00\x00\x0F
user1 column=lf:app2, timestamp=1468906199761, value=\x00\x00\x00\x00\x00\x00\x00\x10
user2 column=lf:app1, timestamp=1468906199806, value=\x00\x00\x00\x00\x00\x00\x00\x11
user2 column=lf:pc1, timestamp=1468906200773, value=\x00\x00\x00\x00\x00\x00\x00\x12
hive> select * from user_app_cookie_list;
OK
user1 15 16
user2 17 NULL
-- 用户userX,作为rowkey,在什么时间(timestamp)作为列名,访问了什么页面的id作为value:page_id (整型变量)
put 'zmtest1', 'user1', 'cf:1399999999', "\x00\x00\x00\x09"
put 'zmtest1', 'user1', 'cf:1400000000', "\x00\x00\x00\x08"
put 'zmtest1', 'user1', 'cf:1400000001', "\x00\x00\x00\x07"
put 'zmtest1', 'user1', 'cf:1400000002', "\x00\x00\x20\xFB"
put 'zmtest1', 'user2', 'cf:1500000000', "\x00\x00\x00\x11"
put 'zmtest1', 'user2', 'cf:1500000001', "\x00\x00\x20\xFC"
表 rowkey 列族:列名 列值
-- hive hbase mapping cf with binary --
http://www.abcn.net/2013/11/hive-hbase-mapping-column-family-with-binary-value.html
CREATE EXTERNAL TABLE ts_string ( username STRING, visits map<string, int> )
STORED BY 'org.apache.hadoop.hive.hbase.HBaseStorageHandler'
WITH SERDEPROPERTIES ("hbase.columns.mapping" = ":key, cf:#s:b")
TBLPROPERTIES("hbase.table.name" = "zmtest1");
CREATE EXTERNAL TABLE ts_int ( username STRING, visits map<int, int> )
STORED BY 'org.apache.hadoop.hive.hbase.HBaseStorageHandler'
WITH SERDEPROPERTIES ("hbase.columns.mapping" = ":key, cf:#s:b")
TBLPROPERTIES("hbase.table.name" = "zmtest1");
CREATE EXTERNAL TABLE ts_int_long ( username STRING, visits map<int, bigint> )
STORED BY 'org.apache.hadoop.hive.hbase.HBaseStorageHandler'
WITH SERDEPROPERTIES ("hbase.columns.mapping" = ":key, cf:#s:b")
TBLPROPERTIES("hbase.table.name" = "zmtest1");
select * from ts_int
lateral view explode(visits) t as ts, page;
select username, ts, page_id from ts_int
lateral view explode(visits) t as ts, page_id;
select username, pos, ts, page_id from ts_int
lateral view posexplode(visits) t as pos, ts, page_id;
username pos ts page_id
user1 1 1399999999 9
user1 2 1400000000 8
user1 3 1400000001 7
user1 4 1400000002 8443
user2 1 1500000000 17
user2 2 1500000001 8444
select username, from_unixtime(ts), page_id from ts_int lateral view explode(visits) t as ts, page_id;