JarCastingbig-sequence-file
a big file, similar to java embedded implementation of kafka, providing sequential data access.
QuickStart
- Sample Code
BigSequenceFile bsf = null;
try {
bsf = new BigSequenceFile("hello.bsf");
bsf.open();
bsf.appendTrunk("abc".getBytes());
bsf.appendTrunk("def".getBytes());
bsf.appendTrunk("g".getBytes());
byte[] trunk = bsf.deductTrunk();
System.out.println(new String(trunk));
} finally {
if (bsf != null) {
bsf.close();
}
}
for more infomation, please read QuickStart.java example.
- maven dependency
<dependency>
<groupId>com.github.downgoon</groupId>
<artifactId>big-sequence-file</artifactId>
<version>0.1.0</version>
</dependency>
- underlying structure
when we new two '.bsf' files: new BigSequenceFile("hello.bsf") and new BigSequenceFile("world.bsf"), the underlying files may look like as follows:
$ tree .
├── hello.bsf
├── hello_0.seg
├── world.bsf
├── world_0.seg
├── world_1.seg
└── world_2.seg
the .bsf file manages meta info of the user namespaced bsf file (e.g. hello.bsf), while multipule .seg files store data info. in general, a BSF file always consists of only one .bsf and several .seg files in underlying storage layer.