通过hive，或者sparksql创建的表，无法实现对单条数据的update和delete。但是hbase可以。对表的操作请参考：hbase 创建表 增删列
一,插入数据
1，插入格式
put ’<table name>’,’<rowkey>’,’<colfamily:colname>’,’<value>’
2，样例
hbase(main):011:0* put 'test_ns:user','1000320190926','login:username','test2' 0 row(s) in 0.0070 seconds hbase(main):033:0> get 'test_ns:user',1000320190926,'login:username' COLUMN CELL login:username timestamp=1569461617502, value=test2 1 row(s) in 0.0300 seconds
二，更新数据
1，更新格式
put ’<table name>’,’<rowkey>’,’<colfamily:colname>’,’<newvalue>’
2，样例
hbase(main):034:0> put 'test_ns:user','1000320190926','login:username','newtest2' 0 row(s) in 0.0100 seconds hbase(main):035:0> get 'test_ns:user',1000320190926,'login:username' COLUMN CELL login:username timestamp=1569463065916, value=newtest2 1 row(s) in 0.0080 seconds
三，get获取数据
1，get获取数据格式
get ’<table name>’,’<rowkey>’,’<colfamily:colname>’,’<colfamily>’,....
get ’<table name>’,’<rowkey>’,’<colfamily>’,{表达式}....
2，样例
//获取user，rowkey为1000320190926，login列族下的所有数据
hbase(main):037:0> get 'test_ns:user',1000320190926,'login'
COLUMN CELL
login:password timestamp=1569461617521, value=pass2
login:username timestamp=1569463065916, value=newtest2
1 row(s) in 0.0070 seconds
//获取user，rowkey为1000320190926，login列族下的所有数据，以及contact列族下tel列的所有数据
hbase(main):038:0> get 'test_ns:user',1000320190926,'login','contact:tel'
COLUMN CELL
contact:tel timestamp=1569461617594, value=02132345678
login:password timestamp=1569461617521, value=pass2
login:username timestamp=1569463065916, value=newtest2
1 row(s) in 0.0060 seconds
//获取user，rowkey为1000320190926，login,info,contact列族下的所有数据，时间范围1569161617594, 1569463065916，版本为1
hbase(main):049:0> get 'test_ns:user',1000320190926, {COLUMN => ['login','info','contact'], \
hbase(main):050:1* TIMERANGE => [1569161617594, 1569463065916], VERSIONS => 1}
COLUMN CELL
contact:mobile timestamp=1569461617577, value=15822345678
contact:tel timestamp=1569461617594, value=02132345678
info:age timestamp=1569461617561, value=36
info:sex timestamp=1569461617538, value=male2
login:password timestamp=1569461617521, value=pass2
1 row(s) in 0.0150 seconds
//获取user，rowkey为1000320190926，login,info列族下的所有数据，contact列族下tel列的所有，时间范围1569161617594, 1569463065916，版本为1
hbase(main):051:0> get 'test_ns:user',1000320190926, {COLUMN => ['login','info','contact:tel'], \
hbase(main):052:1* TIMERANGE => [1569161617594, 1569463065916], VERSIONS => 1}
COLUMN CELL
contact:tel timestamp=1569461617594, value=02132345678
info:age timestamp=1569461617561, value=36
info:sex timestamp=1569461617538, value=male2
login:password timestamp=1569461617521, value=pass2
1 row(s) in 0.0100 seconds
四，scan获取数据
1，scan获取数据格式
scan ’<table name>’,’<colfamily>’,{表达式}....
2，样例
//获取全user表数据
scan 'test_ns:user'
//获取user表，contact列族下tel列的所有数据
scan 'test_ns:user', {COLUMNS => 'contact:tel'}
//获取user表，login,info列族下的2条数据
scan 'test_ns:user', {COLUMNS => ['login', 'info'], LIMIT => 2}
//反向显示表user的数据
scan 'test_ns:user', {REVERSED => true}
//显示所有版本为1的数据，包括删除的数据
scan 'test_ns:user', {RAW => true, VERSIONS => 1}
//获取user表，login,info列族，contact列族tel列，根据时间范围，rowkey的范围获取2条数据
hbase(main):056:0> scan 'test_ns:user', {COLUMNS => ['login','info','contact:tel'], LIMIT => 2, \
hbase(main):057:1* TIMERANGE => [1569401820233, 1569461709184],STARTROW => '1000120190925',ENDROW=> '1000520190925'}
ROW COLUMN+CELL
1000120190925 column=contact:tel, timestamp=1569401821139, value=02112345678
1000220190926 column=contact:tel, timestamp=1569461617469, value=02122345678
1000220190926 column=info:age, timestamp=1569461617408, value=35
1000220190926 column=info:sex, timestamp=1569461617380, value=male1
1000220190926 column=login:password, timestamp=1569461617362, value=pass1
1000220190926 column=login:username, timestamp=1569461617333, value=test1
2 row(s) in 0.0160 seconds
get和scan取数据最大的区别在于，单条，还是多条。get取数据rowkey是必须参数，rowkey是表里面的唯一字段。查询的操作很多，后会单独，在写一篇。
五，append拼接数据
1，append拼接数据格式
append ’<table name>’,’<rowkey>’,’<colfamily:colname>’,’<colfamily>’,....
2，样例
hbase(main):061:0> append 'test_ns:user','1000120190925','login:username','_111' CURRENT VALUE = Tank_111 0 row(s) in 0.0460 seconds
六，delete删除数据
1，格式
delete ‘<table name>’, ‘<rowkey>’, ‘<column name >’, ‘<timestamp>’
2，样例
hbase(main):068:0> get 'test_ns:user','1000120190925','login:username' COLUMN CELL login:username timestamp=1569399395200, value=Tank 1 row(s) in 0.0100 seconds //删除user表下,rowkey为1000120190925，时间为1569399395200，username列数据 hbase(main):069:0> delete 'test_ns:user','1000120190925','login:username',1569399395200 0 row(s) in 0.0070 seconds hbase(main):070:0> get 'test_ns:user','1000120190925','login:username' COLUMN CELL 0 row(s) in 0.0070 seconds
七，deleteall删除数据
1，格式
deleteall ‘<table name>’, ‘<rowkey>’,
2，样例
//删除表user下rowkey为1000120190925所有数据 hbase(main):082:0> deleteall 'test_ns:user','1000120190925' 0 row(s) in 0.0120 seconds //删除表user下rowkey为1000220190926，login列族所有数据 hbase(main):085:0> deleteall 'test_ns:user','1000220190926','login' 0 row(s) in 0.0050 seconds //删除表user下rowkey为1000220190926，info列族下sex列所有数据 hbase(main):086:0> deleteall 'test_ns:user','1000220190926','info:sex' 0 row(s) in 0.0070 seconds //删除表user下rowkey为1000220190926，contact列族所有数据，时间为1569461617380 hbase(main):087:0> deleteall 'test_ns:user','1000220190926','contact',1569461617380 0 row(s) in 0.0060 seconds
八，求行数
hbase(main):089:0> count 'test_ns:user' 4 row(s) in 0.0300 seconds => 4 、、查询表t1中的行数，每10条显示一次，缓存区为1000 hbase(main):090:0> count 'test_ns:user', INTERVAL => 10, CACHE => 1000 4 row(s) in 0.0130 seconds => 4
转载请注明
作者:海底苍鹰
地址:http://blog.51yip.com/hadoop/2188.html