HBase表
每个Table由若干个Region组成,每个Region由若干个Store组成(每个列组构成一个Store),每个由一个MemStore和若干个StoreFiles(HFiles)组成,每个StoreFile由若干个Blocks组成。
HBase对表进行split有三种方式
Pre-splitting,Auto splitting,Forced Splits
本文重点说说Pre-splitting的使用。
Pre-splitting
RegionSplitter提供三个用于预分割的工具:HexStringSplit、SplitAlgorithm、UniformSplit。其中HexStringSplit和UniformSplit是两个预定义的静态类,可以直接使用;而SplitAlgorithm是一个接口,需要开发人员自己实现相应的分隔策略。如果是以十六进制字符串作为行键rowkey或者行键rowkey的前缀是十六进制字符串,用HexStringSplit就比较合适;UniformSplit会把行键均匀地分割多个部分,如果行将rowkey是随机的字节数组,用UniformSplit就比较合适;或者开发者根据需要实现分割策略
1 HexStringSplit
create 'hexstringtb', 'f1', {NUMREGIONS => 15, SPLITALGO => 'HexStringSplit'}
1.1 分区效果
1.2 插入记录
put 'hexstringtb','22222222','f1:name','2'
put 'hexstringtb','55555555','f1:name','5'
1.3 落入分区
22222222-33333333 1条
55555555-66666666 1条
说明 StartKey ≤ RowKey < EndKey
2 UniformSplit
create 'uniformtb', 'f1', {NUMREGIONS => 15, SPLITALGO => 'UniformSplit'}
2.1 分区效果
2.2 插入记录
put 'uniformtb','DDDDDDDD','f1:name','2'
put 'uniformtb','ffffffff','f1:name','5'
2.3 落入分区
DDDDDDDD-UUUUUUUU 1条
ffffffff-wwwwwwww 1条
3 自定义分区
create 'mysplittb', 'f1', SPLITS => ['00|','10|', '20|', '30|', '40|','50|', '60|', '70|', '80|','90|']
3.1 分区效果
3.2 插入记录
put 'mysplittb','00|abcdefd','f1:name','2'
put 'mysplittb','20|abcdefd','f1:name','2'
3.3 落入分区
00|-10| 1条
20|-30| 1条
JavaApi 打印分区信息
/**
* 打印分区信息
* @param tableName
*/
public static boolean printHTableRegionInfo(String tableName) {
try {
Connection connection = getConnection();
RegionLocator r = connection.getRegionLocator(TableName.valueOf(tableName));
List<HRegionLocation> list = r.getAllRegionLocations();
for (HRegionLocation location : list) {
HRegionInfo rg = location.getRegionInfo();
String regionname = Bytes.toString(rg.getRegionName());
String strkey = Bytes.toString(rg.getStartKey());
String endkey = Bytes.toString(rg.getEndKey());
System.out.println(regionname);
System.out.println(strkey);
System.out.println(endkey);
}
return true;
} catch (IOException e) {
log.error("printHTableRegionInfo", e);
}
return false;
}
@Test
public void printHTableRegionInfo(){
String tbName="mysplittb";
if(HBaseClientUtil.printHTableRegionInfo(tbName)){
System.out.println("打印分区信息成功");
}
else{
System.out.println("打印分区信息错误");
}
}
mysplittb,,1547390815132.ef207ecf4712436f160b247c17ddbbd4.
00|
mysplittb,00|,1547390815132.fc59155ab9184e8c6fa59a18b26336f4.
00|
10|
mysplittb,10|,1547390815132.0f6c1ec45cfc1da3744c4de6a57d9bb8.
10|
20|
mysplittb,20|,1547390815132.f843d0cbebdf5622ead01b9f60195e08.
20|
30|
mysplittb,30|,1547390815132.906608d2684c2f1b6e7c2fa6110870fb.
30|
40|
mysplittb,40|,1547390815132.a48bc2ccd28f0c82cef0d8341b2fced6.
40|
50|
mysplittb,50|,1547390815132.c29ac762f05b88ac4d55a420cb97ebe8.
50|
60|
mysplittb,60|,1547390815132.46247c4106c45e9be8a315866348413d.
60|
70|
mysplittb,70|,1547390815132.aa80d22e2d6caaaa4267229d3c9639c3.
70|
80|
mysplittb,80|,1547390815132.a95345bb456c672880ee11475f8db2f1.
80|
90|
mysplittb,90|,1547390815132.37a0a89e8e249d8b9c86eb4cf310b773.
90|