1. 建表并填入測試數(shù)據(jù)
- 首先創(chuàng)建表test儒鹿，并指定一個名為t1的列簇
create 'test','t1'
- 隨意填幾條測試數(shù)據(jù)
put 'test','0001','t1:name','zhangsan'
put 'test','0002','t1:name','lisi'
2. 查看HFile
- 查看命令
hbase hfile -v -p -m -f hdfs://localhost:9000/hbase/data/default/test/782d9564125bae4e0037e367626186b4/t1/44bd781271ef445ea5056ca632569611
- 內(nèi)容如下：
Scanning -> hdfs://localhost:9000/hbase/data/default/test/782d9564125bae4e0037e367626186b4/t1/44bd781271ef445ea5056ca632569611
2019-06-01 22:03:36,540 INFO [main] hfile.CacheConfig: Created cacheConfig: CacheConfig:disabled
K: 0001/t1:name/1559371642919/Put/vlen=8/seqid=4 V: zhangsan
Block index size as per heapsize: 392
reader=hdfs://localhost:9000/hbase/data/default/test/782d9564125bae4e0037e367626186b4/t1/44bd781271ef445ea5056ca632569611,
compression=none,
cacheConf=CacheConfig:disabled,
firstKey=0001/t1:name/1559371642919/Put,
lastKey=0001/t1:name/1559371642919/Put,
avgKeyLen=22,
avgValueLen=8,
entries=1,
length=4914
Trailer:
fileinfoOffset=224,
loadOnOpenDataOffset=115,
dataIndexCount=1,
metaIndexCount=0,
totalUncomressedBytes=4822,
entryCount=1,
compressionCodec=NONE,
uncompressedDataIndexSize=35,
numDataIndexLevels=1,
firstDataBlockOffset=0,
lastDataBlockOffset=0,
comparatorClassName=org.apache.hadoop.hbase.KeyValue$KeyComparator,
encryptionKey=NONE,
majorVersion=3,
minorVersion=0
Fileinfo:
BLOOM_FILTER_TYPE = ROW
DELETE_FAMILY_COUNT = \x00\x00\x00\x00\x00\x00\x00\x00
EARLIEST_PUT_TS = \x00\x00\x01k\x11\xCA\xF8'
KEY_VALUE_VERSION = \x00\x00\x00\x01
LAST_BLOOM_KEY = 0001
MAJOR_COMPACTION_KEY = \x00
MAX_MEMSTORE_TS_KEY = \x00\x00\x00\x00\x00\x00\x00\x04
MAX_SEQ_ID_KEY = 6
TIMERANGE = 1559371642919....1559371642919
hfile.AVG_KEY_LEN = 22
hfile.AVG_VALUE_LEN = 8
hfile.CREATE_TIME_TS = \x00\x00\x01k\x12\x05\xCAt
hfile.LASTKEY = \x00\x040001\x02t1name\x00\x00\x01k\x11\xCA\xF8'\x04
Mid-key: \x00\x040001\x02t1name\x00\x00\x01k\x11\xCA\xF8'\x04
Bloom filter:
BloomSize: 2
No of Keys in bloom: 1
Max Keys for bloom: 1
Percentage filled: 100%
Number of chunks: 1
Comparator: RawBytesComparator
Delete Family Bloom filter:
Not present
Scanned kv count -> 1
注：可以看到第三行K: 0001/t1:name/1559371642919/Put/vlen=8/seqid=4 V: zhangsan葫慎，只有一條數(shù)據(jù)。
原因：此時第二條數(shù)據(jù)在MemStore中，還未Flush到磁盤中氨鹏，因此看不到第二條記錄
- 手動將MemStore數(shù)據(jù)Flush到磁盤
進入hbase shell
flush 'test'
# 查看幫助可以知注竿，flush可以有三種用法
# help "flush"
# hbase> flush 'TABLENAME'
# hbase> flush 'REGIONNAME'
# hbase> flush 'ENCODED_REGIONNAME'
-
然后查看HBase的Web頁面可以看到形成了新的HFile文件
上面的文件是之前查看的文件，下面的是新flush生成的文件
HFile界面.jpg - 用
hbase hfile查看新生成的文件
可以看到第三行是RowKey為0002的那條數(shù)據(jù)
Scanning -> hdfs://localhost:9000/hbase/data/default/test/782d9564125bae4e0037e367626186b4/t1/4aa0b75c76004c8e8a9c5c00819b8611
2019-06-01 22:24:09,582 INFO [main] hfile.CacheConfig: Created cacheConfig: CacheConfig:disabled
K: 0002/t1:name/1559397766755/Put/vlen=4/seqid=9 V: lisi
Block index size as per heapsize: 392
reader=hdfs://localhost:9000/hbase/data/default/test/782d9564125bae4e0037e367626186b4/t1/4aa0b75c76004c8e8a9c5c00819b8611,
compression=none,
cacheConf=CacheConfig:disabled,
firstKey=0002/t1:name/1559397766755/Put,
lastKey=0002/t1:name/1559397766755/Put,
avgKeyLen=22,
avgValueLen=4,
entries=1,
length=4910
Trailer:
fileinfoOffset=220,
loadOnOpenDataOffset=111,
dataIndexCount=1,
metaIndexCount=0,
totalUncomressedBytes=4818,
entryCount=1,
compressionCodec=NONE,
uncompressedDataIndexSize=35,
numDataIndexLevels=1,
firstDataBlockOffset=0,
lastDataBlockOffset=0,
comparatorClassName=org.apache.hadoop.hbase.KeyValue$KeyComparator,
encryptionKey=NONE,
majorVersion=3,
minorVersion=0
Fileinfo:
BLOOM_FILTER_TYPE = ROW
DELETE_FAMILY_COUNT = \x00\x00\x00\x00\x00\x00\x00\x00
EARLIEST_PUT_TS = \x00\x00\x01k\x13Y\x96c
KEY_VALUE_VERSION = \x00\x00\x00\x01
LAST_BLOOM_KEY = 0002
MAJOR_COMPACTION_KEY = \x00
MAX_MEMSTORE_TS_KEY = \x00\x00\x00\x00\x00\x00\x00\x09
MAX_SEQ_ID_KEY = 11
TIMERANGE = 1559397766755....1559397766755
hfile.AVG_KEY_LEN = 22
hfile.AVG_VALUE_LEN = 4
hfile.CREATE_TIME_TS = \x00\x00\x01k\x13a\xE4\xFF
hfile.LASTKEY = \x00\x040002\x02t1name\x00\x00\x01k\x13Y\x96c\x04
Mid-key: \x00\x040002\x02t1name\x00\x00\x01k\x13Y\x96c\x04
Bloom filter:
BloomSize: 2
No of Keys in bloom: 1
Max Keys for bloom: 1
Percentage filled: 100%
Number of chunks: 1
Comparator: RawBytesComparator
Delete Family Bloom filter:
Not present
Scanned kv count -> 1
注：可以使用major_compact "test"命令進行一次全壓縮康嘉，進而合并兩個HFile文件碉输，然后新的HFile中就包含了之前兩條數(shù)據(jù)
