7.1 sort: Sort text files

sort sorts, merges, or compares all the lines from the given files, or standard input if none are given or for a file of ‘-’. By default, sort writes the results to standard output. Synopsis:

sort [option]… [file]…

Many options affect how sort compares lines; if the results are unexpected, try the --debug option to see what happened. A pair of lines is compared as follows: sort compares each pair of fields (see --key), in the order specified on the command line, according to the associated ordering options, until a difference is found or no fields are left. If no key fields are specified, sort uses a default key of the entire line. Finally, as a last resort when all keys compare equal, sort compares entire lines as if no ordering options other than --reverse (-r) were specified. The --stable (-s) option disables this last-resort comparison so that lines in which all fields compare equal are left in their original relative order. The --unique (-u) option also disables the last-resort comparison.

Unless otherwise specified, all comparisons use the character collating sequence specified by the LC_COLLATE locale.2 A line’s trailing newline is not part of the line for comparison purposes. If the final byte of an input file is not a newline, GNU sort silently supplies one. GNU sort (as specified for all GNU utilities) has no limit on input line length or restrictions on bytes allowed within lines.

sort has three modes of operation: sort (the default), merge, and check for sortedness. The following options change the operation mode:


Check whether the given file is already sorted: if it is not all sorted, print a diagnostic containing the first out-of-order line and exit with a status of 1. Otherwise, exit successfully. At most one input file can be given.


Exit successfully if the given file is already sorted, and exit with status 1 otherwise. At most one input file can be given. This is like -c, except it does not print a diagnostic.


Merge the given files by sorting them as a group. Each input file must always be individually sorted. It always works to sort instead of merge; merging is provided because it is faster, in the case where it works.

Exit status:

0 if no error occurred
1 if invoked with -c or -C and the input is not sorted
2 if an error occurred

If the environment variable TMPDIR is set, sort uses its value as the directory for temporary files instead of /tmp. The --temporary-directory (-T) option in turn overrides the environment variable.

The following options affect the ordering of output lines. They may be specified globally or as part of a specific key field. If no key fields are specified, global options apply to comparison of entire lines; otherwise the global options are inherited by key fields that do not specify any special options of their own. In pre-POSIX versions of sort, global options affect only later key fields, so portable shell scripts should specify global options first.


Ignore leading blanks when finding sort keys in each line. By default a blank is a space or a tab, but the LC_CTYPE locale can change this. Note blanks may be ignored by your locale’s collating rules, but without this option they will be significant for character positions specified in keys with the -k option.


Sort in phone directory order: ignore all characters except letters, digits and blanks when sorting. By default letters and digits are those of ASCII and a blank is a space or a tab, but the LC_CTYPE locale can change this.


Fold lowercase characters into the equivalent uppercase characters when comparing so that, for example, ‘b’ and ‘B’ sort as equal. The LC_CTYPE locale determines character types. When used with --unique those lower case equivalent lines are thrown away. (There is currently no way to throw away the upper case equivalent instead. (Any --reverse given would only affect the final result, after the throwing away.))


Sort numerically, converting a prefix of each line to a long double-precision floating point number. See Floating point. Do not report overflow, underflow, or conversion errors. Use the following collating sequence:

  • Lines that do not start with numbers (all considered to be equal).
  • NaNs (“Not a Number” values, in IEEE floating point arithmetic) in a consistent but machine-dependent order.
  • Minus infinity.
  • Finite numbers in ascending numeric order (with -0 and +0 equal).
  • Plus infinity.

Use this option only if there is no alternative; it is much slower than --numeric-sort (-n) and it can lose information when converting to floating point.


Sort numerically, first by numeric sign (negative, zero, or positive); then by SI suffix (either empty, or ‘k’ or ‘K’, or one of ‘MGTPEZY’, in that order; see Block size); and finally by numeric value. For example, ‘1023M’ sorts before ‘1G’ because ‘M’ (mega) precedes ‘G’ (giga) as an SI suffix. This option sorts values that are consistently scaled to the nearest suffix, regardless of whether suffixes denote powers of 1000 or 1024, and it therefore sorts the output of any single invocation of the df, du, or ls commands that are invoked with their --human-readable or --si options. The syntax for numbers is the same as for the --numeric-sort option; the SI suffix must immediately follow the number. Note also the numfmt command, which can be used to reformat numbers to human format after the sort, thus often allowing sort to operate on more accurate numbers.


Ignore nonprinting characters. The LC_CTYPE locale determines character types. This option has no effect if the stronger --dictionary-order (-d) option is also given.


An initial string, consisting of any amount of blanks, followed by a month name abbreviation, is folded to UPPER case and compared in the order ‘JAN’ < ‘FEB’ < … < ‘DEC’. Invalid names compare low to valid names. The LC_TIME locale category determines the month spellings. By default a blank is a space or a tab, but the LC_CTYPE locale can change this.


Sort numerically. The number begins each line and consists of optional blanks, an optional ‘-’ sign, and zero or more digits possibly separated by thousands separators, optionally followed by a decimal-point character and zero or more digits. An empty number is treated as ‘0’. The LC_NUMERIC locale specifies the decimal-point character and thousands separator. By default a blank is a space or a tab, but the LC_CTYPE locale can change this.

Comparison is exact; there is no rounding error.

Neither a leading ‘+’ nor exponential notation is recognized. To compare such strings numerically, use the --general-numeric-sort (-g) option.


Sort by version name and number. It behaves like a standard sort, except that each sequence of decimal digits is treated numerically as an index/version number. (See Version sort ordering.)


Reverse the result of comparison, so that lines with greater key values appear earlier in the output instead of later.


Sort by hashing the input keys and then sorting the hash values. Choose the hash function at random, ensuring that it is free of collisions so that differing keys have differing hash values. This is like a random permutation of the inputs (see shuf invocation), except that keys with the same value sort together.

If multiple random sort fields are specified, the same random hash function is used for all fields. To use different random hash functions for different fields, you can invoke sort more than once.

The choice of hash function is affected by the --random-source option.

Other options are:


Compress any temporary files with the program prog.

With no arguments, prog must compress standard input to standard output, and when given the -d option it must decompress standard input to standard output.

Terminate with an error if prog exits with nonzero status.

White space and the backslash character should not appear in prog; they are reserved for future use.


Disallow processing files named on the command line, and instead process those named in file file; each name being terminated by a zero byte (ASCII NUL). This is useful when the list of file names is so long that it may exceed a command line length limitation. In such cases, running sort via xargs is undesirable because it splits the list into pieces and makes sort print sorted output for each sublist rather than for the entire list. One way to produce a list of ASCII NUL terminated file names is with GNU find, using its -print0 predicate. If file is ‘-’ then the ASCII NUL terminated file names are read from standard input.

-k pos1[,pos2]

Specify a sort field that consists of the part of the line between pos1 and pos2 (or the end of the line, if pos2 is omitted), inclusive.

In its simplest form pos specifies a field number (starting with 1), with fields being separated by runs of blank characters, and by default those blanks being included in the comparison at the start of each field. To adjust the handling of blank characters see the -b and -t options.

More generally, each pos has the form ‘f[.c][opts]’, where f is the number of the field to use, and c is the number of the first character from the beginning of the field. Fields and character positions are numbered starting with 1; a character position of zero in pos2 indicates the field’s last character. If ‘.c’ is omitted from pos1, it defaults to 1 (the beginning of the field); if omitted from pos2, it defaults to 0 (the end of the field). opts are ordering options, allowing individual keys to be sorted according to different rules; see below for details. Keys can span multiple fields.

Example: To sort on the second field, use --key=2,2 (-k 2,2). See below for more notes on keys and more examples. See also the --debug option to help determine the part of the line being used in the sort.


Highlight the portion of each line used for sorting. Also issue warnings about questionable usage to stderr.


Merge at most nmerge inputs at once.

When sort has to merge more than nmerge inputs, it merges them in groups of nmerge, saving the result in a temporary file, which is then used as an input in a subsequent merge.

A large value of nmerge may improve merge performance and decrease temporary storage utilization at the expense of increased memory usage and I/O. Conversely a small value of nmerge may reduce memory requirements and I/O at the expense of temporary storage consumption and merge performance.

The value of nmerge must be at least 2. The default value is currently 16, but this is implementation-dependent and may change in the future.

The value of nmerge may be bounded by a resource limit for open file descriptors. The commands ‘ulimit -n’ or ‘getconf OPEN_MAX’ may display limits for your systems; these limits may be modified further if your program already has some files open, or if the operating system has other limits on the number of open files. If the value of nmerge exceeds the resource limit, sort silently uses a smaller value.

-o output-file

Write output to output-file instead of standard output. Normally, sort reads all input before opening output-file, so you can sort a file in place by using commands like sort -o F F and cat F | sort -o F. However, it is often safer to output to an otherwise-unused file, as data may be lost if the system crashes or sort encounters an I/O or other serious error while a file is being sorted in place. Also, sort with --merge (-m) can open the output file before reading all input, so a command like cat F | sort -m -o F - G is not safe as sort might start writing F before cat is done reading it.

On newer systems, -o cannot appear after an input file if POSIXLY_CORRECT is set, e.g., ‘sort F -o F’. Portable scripts should specify -o output-file before any input files.


Use file as a source of random data used to determine which random hash function to use with the -R option. See Random sources.


Make sort stable by disabling its last-resort comparison. This option has no effect if no fields or global ordering options other than --reverse (-r) are specified.

-S size

Use a main-memory sort buffer of the given size. By default, size is in units of 1024 bytes. Appending ‘%’ causes size to be interpreted as a percentage of physical memory. Appending ‘K’ multiplies size by 1024 (the default), ‘M’ by 1,048,576, ‘G’ by 1,073,741,824, and so on for ‘T’, ‘P’, ‘E’, ‘Z’, and ‘Y’. Appending ‘b’ causes size to be interpreted as a byte count, with no multiplication.

This option can improve the performance of sort by causing it to start with a larger or smaller sort buffer than the default. However, this option affects only the initial buffer size. The buffer grows beyond size if sort encounters input lines larger than size.

-t separator

Use character separator as the field separator when finding the sort keys in each line. By default, fields are separated by the empty string between a non-blank character and a blank character. By default a blank is a space or a tab, but the LC_CTYPE locale can change this.

That is, given the input line ‘ foo bar’, sort breaks it into fields ‘ foo’ and ‘ bar’. The field separator is not considered to be part of either the field preceding or the field following, so with ‘sort -t " "’ the same input line has three fields: an empty field, ‘foo’, and ‘bar’. However, fields that extend to the end of the line, as -k 2, or fields consisting of a range, as -k 2,3, retain the field separators present between the endpoints of the range.

To specify ASCII NUL as the field separator, use the two-character string ‘\0’, e.g., ‘sort -t '\0'’.

-T tempdir

Use directory tempdir to store temporary files, overriding the TMPDIR environment variable. If this option is given more than once, temporary files are stored in all the directories given. If you have a large sort or merge that is I/O-bound, you can often improve performance by using this option to specify directories on different disks and controllers.


Set the number of sorts run in parallel to n. By default, n is set to the number of available processors, but limited to 8, as there are diminishing performance gains after that. Note also that using n threads increases the memory usage by a factor of log n. Also see nproc invocation.


Normally, output only the first of a sequence of lines that compare equal. For the --check (-c or -C) option, check that no pair of consecutive lines compares equal.

This option also disables the default last-resort comparison.

The commands sort -u and sort | uniq are equivalent, but this equivalence does not extend to arbitrary sort options. For example, sort -n -u inspects only the value of the initial numeric string when checking for uniqueness, whereas sort -n | uniq inspects the entire line. See uniq invocation.


Delimit items with a zero byte rather than a newline (ASCII LF). I.e., treat input as items separated by ASCII NUL and terminate output items with ASCII NUL. This option can be useful in conjunction with ‘perl -0’ or ‘find -print0’ and ‘xargs -0’ which do the same in order to reliably handle arbitrary file names (even those containing blanks or other special characters).

Historical (BSD and System V) implementations of sort have differed in their interpretation of some options, particularly -b, -f, and -n. GNU sort follows the POSIX behavior, which is usually (but not always!) like the System V behavior. According to POSIX, -n no longer implies -b. For consistency, -M has been changed in the same way. This may affect the meaning of character positions in field specifications in obscure cases. The only fix is to add an explicit -b.

A position in a sort field specified with -k may have any of the option letters ‘MbdfghinRrV’ appended to it, in which case no global ordering options are inherited by that particular field. The -b option may be independently attached to either or both of the start and end positions of a field specification, and if it is inherited from the global options it will be attached to both. If input lines can contain leading or adjacent blanks and -t is not used, then -k is typically combined with -b or an option that implicitly ignores leading blanks (‘Mghn’) as otherwise the varying numbers of leading blanks in fields can cause confusing results.

If the start position in a sort field specifier falls after the end of the line or after the end field, the field is empty. If the -b option was specified, the ‘.c’ part of a field specification is counted from the first nonblank character of the field.

On systems not conforming to POSIX 1003.1-2001, sort supports a traditional origin-zero syntax ‘+pos1 [-pos2]’ for specifying sort keys. The traditional command ‘sort +a.x -b.y’ is equivalent to ‘sort -k a+1.x+1,b’ if y is ‘0’ or absent, otherwise it is equivalent to ‘sort -k a+1.x+1,b+1.y’.

This traditional behavior can be controlled with the _POSIX2_VERSION environment variable (see Standards conformance); it can also be enabled when POSIXLY_CORRECT is not set by using the traditional syntax with ‘-pos2’ present.

Scripts intended for use on standard hosts should avoid traditional syntax and should use -k instead. For example, avoid ‘sort +2’, since it might be interpreted as either ‘sort ./+2’ or ‘sort -k 3’. If your script must also run on hosts that support only the traditional syntax, it can use a test like ‘if sort -k 1 </dev/null >/dev/null 2>&1; then …’ to decide which syntax to use.

Here are some examples to illustrate various combinations of options.

  • Sort in descending (reverse) numeric order.
    sort -n -r
  • Run no more than 4 sorts concurrently, using a buffer size of 10M.
    sort --parallel=4 -S 10M
  • Sort alphabetically, omitting the first and second fields and the blanks at the start of the third field. This uses a single key composed of the characters beginning at the start of the first nonblank character in field three and extending to the end of each line.
    sort -k 3b
  • Sort numerically on the second field and resolve ties by sorting alphabetically on the third and fourth characters of field five. Use ‘:’ as the field delimiter.

    sort -t : -k 2,2n -k 5.3,5.4

    Note that if you had written -k 2n instead of -k 2,2n sort would have used all characters beginning in the second field and extending to the end of the line as the primary numeric key. For the large majority of applications, treating keys spanning more than one field as numeric will not do what you expect.

    Also note that the ‘n’ modifier was applied to the field-end specifier for the first key. It would have been equivalent to specify -k 2n,2 or -k 2n,2n. All modifiers except ‘b’ apply to the associated field, regardless of whether the modifier character is attached to the field-start and/or the field-end part of the key specifier.

  • Sort the password file on the fifth field and ignore any leading blanks. Sort lines with equal values in field five on the numeric user ID in field three. Fields are separated by ‘:’.

    sort -t : -k 5b,5 -k 3,3n /etc/passwd
    sort -t : -n -k 5b,5 -k 3,3 /etc/passwd
    sort -t : -b -k 5,5 -k 3,3n /etc/passwd

    These three commands have equivalent effect. The first specifies that the first key’s start position ignores leading blanks and the second key is sorted numerically. The other two commands rely on global options being inherited by sort keys that lack modifiers. The inheritance works in this case because -k 5b,5b and -k 5b,5 are equivalent, as the location of a field-end lacking a ‘.c’ character position is not affected by whether initial blanks are skipped.

  • Sort a set of log files, primarily by IPv4 address and secondarily by timestamp. If two lines’ primary and secondary keys are identical, output the lines in the same order that they were input. The log files contain lines that look like this: - - [01/Apr/2004:06:31:51 +0000] message 1 - - [24/Apr/2004:20:17:39 +0000] message 2

    Fields are separated by exactly one space. Sort IPv4 addresses lexicographically, e.g., sorts before because 61 is less than 129.

    sort -s -t ' ' -k 4.9n -k 4.5M -k 4.2n -k 4.14,4.21 file*.log |
    sort -s -t '.' -k 1,1n -k 2,2n -k 3,3n -k 4,4n

    This example cannot be done with a single sort invocation, since IPv4 address components are separated by ‘.’ while dates come just after a space. So it is broken down into two invocations of sort: the first sorts by timestamp and the second by IPv4 address. The timestamp is sorted by year, then month, then day, and finally by hour-minute-second field, using -k to isolate each field. Except for hour-minute-second there’s no need to specify the end of each key field, since the ‘n’ and ‘M’ modifiers sort based on leading prefixes that cannot cross field boundaries. The IPv4 addresses are sorted lexicographically. The second sort uses ‘-s’ so that ties in the primary key are broken by the secondary key; the first sort uses ‘-s’ so that the combination of the two sorts is stable.

  • Generate a tags file in case-insensitive sorted order.

    find src -type f -print0 | sort -z -f | xargs -0 etags --append

    The use of -print0, -z, and -0 in this case means that file names that contain blanks or other special characters are not broken up by the sort operation.

  • Use the common DSU, Decorate Sort Undecorate idiom to sort lines according to their length.

    awk '{print length, $0}' /etc/passwd | sort -n | cut -f2- -d' '

    In general this technique can be used to sort data that the sort command does not support, or is inefficient at, sorting directly.

  • Shuffle a list of directories, but preserve the order of files within each directory. For instance, one could use this to generate a music playlist in which albums are shuffled but the songs of each album are played in order.
    ls */* | sort -t / -k 1,1R -k 2,2



If you use a non-POSIX locale (e.g., by setting LC_ALL to ‘en_US’), then sort may produce output that is sorted differently than you’re accustomed to. In that case, set the LC_ALL environment variable to ‘C’. Note that setting only LC_COLLATE has two problems. First, it is ineffective if LC_ALL is also set. Second, it has undefined behavior if LC_CTYPE (or LANG, if LC_CTYPE is unset) is set to an incompatible value. For example, you get undefined behavior if LC_CTYPE is ja_JP.PCK but LC_COLLATE is en_US.UTF-8.