1LZ4(1) User Commands LZ4(1)
2
6 lz4 - lz4, unlz4, lz4cat - Compress or decompress .lz4 files
7
9 lz4 [OPTIONS] [-|INPUT-FILE] OUTPUT-FILE
10 unlz4 is equivalent to lz4 -d
12 lz4cat is equivalent to lz4 -dcfm
14 When writing scripts that need to decompress files, it is recommended
16 to always use the name lz4 with appropriate arguments (lz4 -d or lz4
17 -dc) instead of the names unlz4 and lz4cat.
18
20 lz4 is an extremely fast lossless compression algorithm, based on
21 byte-aligned LZ77 family of compression scheme. lz4 offers compression
22 speeds of 400 MB/s per core, linearly scalable with multi-core CPUs. It
23 features an extremely fast decoder, with speed in multiple GB/s per
24 core, typically reaching RAM speed limit on multi-core systems. The
25 native file format is the .lz4 format.
26 Difference between lz4 and gzip
28 lz4 supports a command line syntax similar but not identical to
29 gzip(1). Differences are :
30 · lz4 preserves original files
32 · lz4 compresses a single file by default (see -m for multiple files)
34 · lz4 file1 file2 means : compress file1 into file2
36 · lz4 file.lz4 will default to decompression (use -z to force com‐
38 pression)
39 · lz4 shows real-time notification statistics during compression or
41 decompression of a single file (use -q to silence them)
42 · If no destination name is provided, result is sent to stdout except
44 if stdout is the console.
45 · If no destination name is provided, and if stdout is the console,
47 file is compressed into file.lz4.
48 · As a consequence of previous rules, note the following example :
50 lz4 file | consumer sends compressed data to consumer through std‐
51 out, hence it does not create file.lz4.
52 · Another consequence of those rules is that to run lz4 under nohup,
54 you should provide a destination file: nohup lz4 file file.lz4,
55 because nohup writes the specified command´s output to a file.
56 Default behaviors can be modified by opt-in commands, detailed below.
60 · lz4 -m makes it possible to provide multiple input filenames, which
62 will be compressed into files using suffix .lz4. Progress notifica‐
63 tions are also disabled by default (use -v to enable them). This
64 mode has a behavior which more closely mimics gzip command line,
65 with the main remaining difference being that source files are pre‐
66 served by default.
67 · Similarly, lz4 -m -d can decompress multiple *.lz4 files.
69 · It´s possible to opt-in to erase source files on successful com‐
71 pression or decompression, using --rm command.
72 · Consequently, lz4 -m --rm behaves the same as gzip.
74 Concatenation of .lz4 files
78 It is possible to concatenate .lz4 files as is. lz4 will decompress
79 such files as if they were a single .lz4 file. For example:
80 lz4 file1 > foo.lz4
84 lz4 file2 >> foo.lz4
85 Then lz4cat foo.lz4 is equivalent to cat file1 file2.
89
91 Short commands concatenation
92 In some cases, some options can be expressed using short command -x or
93 long command --long-word. Short commands can be concatenated together.
94 For example, -d -c is equivalent to -dc. Long commands cannot be con‐
95 catenated. They must be clearly separated by a space.
96 Multiple commands
98 When multiple contradictory commands are issued on a same command line,
99 only the latest one will be applied.
100 Operation mode
102 -z --compress
103 Compress. This is the default operation mode when no operation
104 mode option is specified, no other operation mode is implied
105 from the command name (for example, unlz4 implies --decompress),
106 nor from the input file name (for example, a file extension .lz4
107 implies --decompress by default). -z can also be used to force
108 compression of an already compressed .lz4 file.
109 -d --decompress --uncompress
111 Decompress. --decompress is also the default operation when the
112 input filename has an .lz4 extension.
113 -t --test
115 Test the integrity of compressed .lz4 files. The decompressed
116 data is discarded. No files are created nor removed.
117 -b# Benchmark mode, using # compression level.
119 Operation modifiers
121 -# Compression level, with # being any value from 1 to 12. Higher
122 values trade compression speed for compression ratio. Values
123 above 12 are considered the same as 12. Recommended values are 1
124 for fast compression (default), and 9 for high compression.
125 Speed/compression trade-off will vary depending on data to com‐
126 press. Decompression speed remains fast at all settings.
127 -D dictionaryName
129 Compress, decompress or benchmark using dictionary dictionary‐
130 Name. Compression and decompression must use the same dictionary
131 to be compatible. Using a different dictionary during decompres‐
132 sion will either abort due to decompression error, or generate a
133 checksum error.
134 -f --[no-]force
136 This option has several effects:
137 If the target file already exists, overwrite it without prompt‐
139 ing.
140 When used with --decompress and lz4 cannot recognize the type of
142 the source file, copy the source file as is to standard output.
143 This allows lz4cat --force to be used like cat (1) for files
144 that have not been compressed with lz4.
145 -c --stdout --to-stdout
147 Force write to standard output, even if it is the console.
148 -m --multiple
150 Multiple input files. Compressed file names will be appended a
151 .lz4 suffix. This mode also reduces notification level. lz4 -m
152 has a behavior equivalent to gzip -k (it preserves source files
153 by default).
154 -r operate recursively on directories. This mode also sets -m (mul‐
156 tiple input files).
157 -B# Block size [4-7](default : 7)
159 -B4= 64KB ; -B5= 256KB ; -B6= 1MB ; -B7= 4MB
160 -BD Block Dependency (improves compression ratio on small blocks)
162 --fast[=#]
164 switch to ultra-fast compression levels. If =# is not present,
165 it defaults to 1. The higher the value, the faster the compres‐
166 sion speed, at the cost of some compression ratio. This setting
167 overwrites compression level if one was set previously. Simi‐
168 larly, if a compression level is set after --fast, it overrides
169 it.
170 --[no-]frame-crc
172 Select frame checksum (default:enabled)
173 --[no-]content-size
175 Header includes original size (default:not present)
176 Note : this option can only be activated when the original size
177 can be determined, hence for a file. It won´t work with unknown
178 source size, such as stdin or pipe.
179 --[no-]sparse
181 Sparse mode support (default:enabled on file, disabled on std‐
182 out)
183 -l Use Legacy format (typically for Linux Kernel compression)
185 Note : -l is not compatible with -m (--multiple) nor -r
186 Other options
188 -v --verbose
189 Verbose mode
190 -q --quiet
192 Suppress warnings and real-time statistics; specify twice to
193 suppress errors too
194 -h -H --help
196 Display help/long help and exit
197 -V --version
199 Display Version number and exit
200 -k --keep
202 Preserve source files (default behavior)
203 --rm Delete source files on successful compression or decompression
205 -- Treat all subsequent arguments as files
207 Benchmark mode
209 -b# Benchmark file(s), using # compression level
210 -e# Benchmark multiple compression levels, from b# to e# (included)
212 -i# Minimum evaluation time in seconds [1-9] (default : 3)
214
216 Report bugs at: https://github.com/lz4/lz4/issues
217
219 Yann Collet
220lz4 1.8.3 September 2018 LZ4(1)