lsort - Sort the elements of a list


lsort ?options? list


This command sorts the elements of list, returning a new list in sorted order. The implementation of the lsort command uses the merge-sort algorithm which is a stable sort that has O(n log n) performance characteristics.

By default ASCII sorting is used with the result returned in increasing order. However, any of the following options may be specified before list to control the sorting process (unique abbreviations are accepted):

. Use string comparison with Unicode code-point collation order (the name is for backward-compatibility reasons.) This is the default.
. Use dictionary-style comparison. This is the same as -ascii except (a) case is ignored except as a tie-breaker and (b) if two strings contain embedded numbers, the numbers compare as integers, not characters. For example, in -dictionary mode, bigBoy sorts between bigbang and bigboy, and x10y sorts between x9y and x11y. Overrides the -nocase option.
. Convert list elements to integers and use integer comparison.
. Convert list elements to floating-point values and use floating comparison.
-command command
. Use command as a comparison command. To compare two elements, evaluate a Tcl script consisting of command with the two elements appended as additional arguments. The script should return an integer less than, equal to, or greater than zero if the first element is to be considered less than, equal to, or greater than the second, respectively.
. Sort the list in increasing order (“smallest”) This is the default.
. Sort the list in decreasing order (“largest”)
. Return a list of indices into list in sorted order instead of the values themselves.
-index indexList
. If this option is specified, each of the elements of list must itself be a proper Tcl sublist (unless -stride is used). Instead of sorting based on whole sublists, lsort will extract the indexList'th element from each sublist (as if the overall element and the indexList were passed to lindex) and sort based on the given element. For example,

lsort -integer -index 1 \
      {{First 24} {Second 18} {Third 30}}

returns {Second 18} {First 24} {Third 30},

lsort -index end-1 \
        {{a 1 e i} {b 2 3 f g} {c 4 5 6 d h}}

returns {c 4 5 6 d h} {a 1 e i} {b 2 3 f g}, and

lsort -index {0 1} {
    {{b i g} 12345}
    {{d e m o} 34512}
    {{c o d e} 54321}

returns {{d e m o} 34512} {{b i g} 12345} {{c o d e} 54321} (because e sorts before i which sorts before o.) This option is much more efficient than using -command to achieve the same effect.

-stride strideLength
. If this option is specified, the list is treated as consisting of groups of strideLength elements and the groups are sorted by either their first element or, if the -index option is used, by the element within each group given by the first index passed to -index (which is then ignored by -index). Elements always remain in the same position within their group.

The list length must be an integer multiple of strideLength, which in turn must be at least 2.

For example,

lsort -stride 2 {carrot 10 apple 50 banana 25}

returns “apple 50 banana 25 carrot 10” and

lsort -stride 2 -index 1 -integer {carrot 10 apple 50 banana 25}

returns “carrot 10 banana 25 apple 50”

. Causes comparisons to be handled in a case-insensitive manner. Has no effect if combined with the -dictionary, -integer, or -real options.
. If this option is specified, then only the last set of duplicate elements found in the list will be retained. Note that duplicates are determined relative to the comparison used in the sort. Thus if -index 0 is used, {1 a} and {1 b} would be considered duplicates and only the second element, {1 b}, would be retained.


The options to lsort only control what sort of comparison is used, and do not necessarily constrain what the values themselves actually are. This distinction is only noticeable when the list to be sorted has fewer than two elements.

The lsort command is reentrant, meaning it is safe to use as part of the implementation of a command used in the -command option.


Sorting a list using ASCII sorting:

% lsort {a10 B2 b1 a1 a2}
B2 a1 a10 a2 b1

Sorting a list using Dictionary sorting:

% lsort -dictionary {a10 B2 b1 a1 a2}
a1 a2 a10 b1 B2

Sorting lists of integers:

% lsort -integer {5 3 1 2 11 4}
1 2 3 4 5 11
% lsort -integer {1 2 0x5 7 0 4 -1}
-1 0 1 2 4 0x5 7

Sorting lists of floating-point numbers:

% lsort -real {5 3 1 2 11 4}
1 2 3 4 5 11
% lsort -real {.5 0.07e1 0.4 6e-1}
0.4 .5 6e-1 0.07e1

Sorting using indices:

% # Note the space character before the c
% lsort {{a 5} { c 3} {b 4} {e 1} {d 2}}
{ c 3} {a 5} {b 4} {d 2} {e 1}
% lsort -index 0 {{a 5} { c 3} {b 4} {e 1} {d 2}}
{a 5} {b 4} { c 3} {d 2} {e 1}
% lsort -index 1 {{a 5} { c 3} {b 4} {e 1} {d 2}}
{e 1} {d 2} { c 3} {b 4} {a 5}

Sorting a dictionary:

% set d [dict create c d a b h i f g c e]
c e a b h i f g
% lsort -stride 2 $d
a b c e f g h i

Sorting using striding and multiple indices:

% # Note the first index value is relative to the group
% lsort -stride 3 -index {0 1} \
     {{Bob Smith} 25 Audi {Jane Doe} 40 Ford}
{{Jane Doe} 40 Ford {Bob Smith} 25 Audi}

Stripping duplicate values using sorting:

% lsort -unique {a b c a b c a b c}
a b c

More complex sorting using a comparison function:

% proc compare {a b} {
    set a0 [lindex $a 0]
    set b0 [lindex $b 0]
    if {$a0 < $b0} {
        return -1
    } elseif {$a0 > $b0} {
        return 1
    return [string compare [lindex $a 1] [lindex $b 1]]
% lsort -command compare \
        {{3 apple} {0x2 carrot} {1 dingo} {2 banana}}
{1 dingo} {2 banana} {0x2 carrot} {3 apple}


list, lappend, lindex, linsert, llength, lsearch, lset, lrange, lreplace


element, list, order, sort