# aligned-block-file

**Repository Path**: mirrors_regular/aligned-block-file

## Basic Information

- **Project Name**: aligned-block-file
- **Description**: No description available
- **Primary Language**: Unknown
- **License**: MIT
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2020-09-25
- **Last Updated**: 2026-05-17

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# aligned-block-file

read and write to a file in a cache-friendly way by using aligned blocks.

This module provides an interface to read arbitary buffers from a file,
and manages it as a series of aligned buffers. This allows you to write
high performance binary file formats, where many reads do not necessarily
mean many fs reads.

## Blocks(filename, block_size, flags) => AlignedBlockFile

create an instance, `block_size` is the size of the underlying block.
I suggest 1024 or 4096 or some multiple of your OS block size.
`flags` is passed to [fs.open](http://devdocs.io/node/fs#fs_fs_open_path_flags_mode_callback)

### abf.read(start, end, cb)

read a buffer from the file. If the range is already in the cache
`cb` will be called synchronously.

### abf.readUInt32BE (start, cb)

read a UInt32BE from the file. (`cb` may be sync, if the buffer is already in cache)

### abf.size

an observable of the files size.

### abf.offset

an observable of the end of the file

### append (buf, cb)

append `buf` to the file.

this must not be called again until the previous call has returned.
updated values for size and offset will be triggered immediately before the `cb` is called.

### truncate(length, cb)

shorten the file to `length` removing anything after that point.

## License

MIT