Hbase table DDL操作及scala API操作

Hbase shell操作table

建表

• 直接创建
  ns3 是namespace,emp是表,base_info是列簇

hbase(main):037:0> create 'ns3:emp','base_info'

shell界面会有结果提示,如下,表示操作的结果创建了一个叫做emp的表

=> Hbase::Table - ns3:emp

• 使用变量

  hbase(main):036:0> t = create 'ns3:employee','base_info'

• 创建多个列族名

  hbase(main):040:0> create 'ns3:emp1','f1','f2','f3'

• 创建表时同时指定列族属性

 hbase(main):043:0> create 'ns3:emp2',
  {NAME=>'f1',VERSIONS'10=>3,TTL=>2592000,BLOCKCACHE=>TRUE}

• 创建表同时指定分片

  hbase(main):050:0> create 'ns3:emp=3', 'f1', SPLITS => ['10', '20', '30', '40']

查看表信息

hbase(main):007:0> describe 'ns3:emp1'
Table ns3:emp1 is ENABLED
ns3:emp1
COLUMN FAMILIES DESCRIPTION
{NAME => 'f1', BLOOMFILTER => 'ROW', VERSIONS => '1', IN_MEMORY => 'false', KEE
P_DELETED_CELLS => 'FALSE', DATA_BLOCK_ENCODING => 'NONE', TTL => 'FOREVER', CO
MPRESSION => 'NONE', MIN_VERSIONS => '0', BLOCKCACHE => 'true', BLOCKSIZE => '6
5536', REPLICATION_SCOPE => '0'}

列出所有表

hbase(main):008:0> list
TABLE
AK_POPULATION
SYSTEM:CATALOG
SYSTEM:FUNCTION
SYSTEM:MUTEX
SYSTEM:SEQUENCE
SYSTEM:STATS
Student

修改表的属性

hbase(main):011:0> alter 'csdn:emp',{NAME=>'f1',BLOOMFILTER=>'rowcol'}
Updating all regions with the new schema...
1/1 regions updated.

hbase(main):012:0> describe 'csdn:emp'
COLUMN FAMILIES DESCRIPTION
{NAME => 'base_info', BLOOMFILTER => 'ROW', VERSIONS => '1', IN_MEMORY => 'false', KEEP_DELETED_CELLS => 'FALSE', DATA_BLOCK_ENCODING => 'NONE', TTL => '
FOREVER', COMPRESSION => 'NONE', MIN_VERSIONS => '0', BLOCKCACHE => 'true', BLOCKSIZE => '65536', REPLICATION_SCOPE => '0'}
{NAME => 'f1', BLOOMFILTER => 'ROWCOL', VERSIONS => '1', IN_MEMORY => 'false', KEEP_DELETED_CELLS => 'FALSE', DATA_BLOCK_ENCODING => 'NONE', TTL => 'FORE
VER', COMPRESSION => 'NONE', MIN_VERSIONS => '0', BLOCKCACHE => 'true', BLOCKSIZE => '65536', REPLICATION_SCOPE => '0'}
2 row(s) in 0.1070 seconds

增加列族

//逗号不能省略
hbase(main):013:0> alter 'csdn:emp' {NAME=>'f2'}
SyntaxError: (hbase):13: syntax error, unexpected tLCURLY

alter 'csdn:emp' {NAME=>'f2'}
                 ^

hbase(main):014:0> alter 'csdn:emp', {NAME=>'f2'}
Updating all regions with the new schema...
//用逗号隔开
hbase(main):015:0> alter 'csdn:emp','f3'
Updating all regions with the new schema...
//可以同时增加多个列族
hbase(main):016:0> alter 'csdn:emp','f4','f5'
Updating all regions with the new schema...

删除列族

hbase(main):018:0> alter 'csdn:emp','delete'=>'f1'

删除表
先禁用才能删除,直接删除是删除不了的~

hbase(main):003:0> disable 'csdn:emp'
0 row(s) in 2.4470 seconds

hbase(main):004:0> drop 'csdn:emp'
0 row(s) in 1.3730 seconds

Hbase table DDL scala编程

首先创建工具类备用

object HbaseUtil {
  private val conf = new Configuration()
  conf.set("hbase.zookeeper.quorum", "mypc01:2181,mypc02:2181,mypc03:2181")
  private val connection: Connection = ConnectionFactory.createConnection(conf)
  var admin: Admin = _
  val table: Table = null

  def getAdmin: Admin = {
    admin = connection.getAdmin
    admin
  }

  def closeAdmin(): Unit = {
    admin.close()
  }

  def getTable(name: String): Table = {
     //获取tableName对象
    val tableName: TableName = TableName.valueOf(name)
    //从连接获取Table对象
    val table: Table = connection.getTable(tableName)
    table
  }

  def closeTable(): Unit = {
    table.close()
  }
}

创建表格

object CreateTable extends App {
  private val name: TableName = TableName.valueOf("csdn:students")
  //新建一个表描述符
  private val tabldedesc = new HTableDescriptor(name)
  //新建一个列族描述符
  private val cf = new HColumnDescriptor("f1")
  //对列族设置布隆过滤器
  cf.setBloomFilterType(BloomType.ROWCOL)
  cf.setInMemory(true)
  //设置列族保存的版本数
  cf.setVersions(1,3)
  //设置列族单元格的过期时间
  cf.setTimeToLive(24*60*60)
  //列族绑定表
  tabldedesc.addFamily(cf)
  //admin创建表格
  HbaseUtil.getAdmin.createTable(tabldedesc)
}

修改表的列族信息

上面的列族,表都是new的,这里是需要获取!

object AlterCF extends App {
  private val name: TableName = TableName.valueOf("csdn:students")
  //利用admin对象获取表描述符
  private val tdescriptor: HTableDescriptor = HbaseUtil.getAdmin.getTableDescriptor(name)
  //从表对象上获取列族描述符
  private val hdescriptor: HColumnDescriptor = tdescriptor.getFamily(Bytes.toBytes("f1"))
  //修改列族
  hdescriptor.setBloomFilterType(BloomType.ROW)
  hdescriptor.setTimeToLive(24*60)
  //提交修改
  HbaseUtil.getAdmin.modifyColumn(name,hdescriptor)
}

向表中添加新的列族

新建一个列族调用admin对象的addColumn方法就可以了.这里的column实际指的是CF,就是列族

object AddCF extends App {
  private val tableName: TableName = TableName.valueOf("csdn:students")
  private val tdescriptor: HTableDescriptor = HbaseUtil.getAdmin.getTableDescriptor(tableName)
  private val hdescriptor = new HColumnDescriptor("f2")
  hdescriptor.setMaxVersions(5)
  HbaseUtil.getAdmin.addColumn(tableName,hdescriptor)
}

获取表的描述信息

其实是获取列族的描述信息

object TabldeDesc extends App {
  private val tableName: TableName = TableName.valueOf("csdn:students")
  private val tdescriptor: HTableDescriptor = HbaseUtil.getAdmin.getTableDescriptor(tableName)
  private val families: Array[HColumnDescriptor] = tdescriptor.getColumnFamilies
  for(x<-families){
    println(s"cf: ${x.getNameAsString} blommfilter: ${x.getBloomFilterType} live: ${x.getTimeToLive}")
  }
}

删除表格

要先禁用才能删除表格~

object deleteTable extends App {
  private val name: TableName = TableName.valueOf("csdn:students")
  if(HbaseUtil.getAdmin.tableExists(name)){
    if(!HbaseUtil.getAdmin.isTableDisabled(name)) HbaseUtil.getAdmin.disableTable(name)
  }
  HbaseUtil.getAdmin.deleteTable(name)
}

总结

• 表格的DDL操作,其实主要是表和列族的相关操作,以及列族的属性是比较丰富的,可以修改!
• 认清三大对象,Admin,HTableDescriptor对象以及HColumnDescriptor对象,认识这三个对象,其他都好办了~