proc array:addn {_a key value} { upvar $_a a set last [lindex [lsort [array names a]] end] set no [lindex [split $last ,] 0] if {![string length $no]} {set no 0} incr no set a([format %06.6d $no],$key) $value } array:addn foo bar testing testing array:addn foo baz testing2 testing2 foreach i [lsort [array names foo]] {puts $i:$foo($i)} 000001,bar:testing ~ 000002,baz:testing2 #... and here's a quick-and dirty retriever for a keyed element: proc array:getn {_a key} { upvar $_a a set a([array names a *,$key]) } array:getn foo bar testing array:getn foo barry can't read "a()": no such element in array
Stephen Trier - The wonderful -dictionary option to lsort solves the millionth-element problem:
proc array:addn {_a key value} { upvar $_a a set last [lindex [lsort -dictionary [array names a]] end] set no [lindex [split $last ,] 0] if {![string length $no]} {set no 0} incr no set a($no,$key) $value } foreach x {foo bar a b c d e f g h i j k l} { array:addn foo $x "testing $x" } foreach i [lsort -dictionary [array names foo]] {puts $i:$foo($i)} 1,foo:testing foo 2,bar:testing bar 3,a:testing a 4,b:testing b 5,c:testing c 6,d:testing d 7,e:testing e 8,f:testing f 9,g:testing g 10,h:testing h 11,i:testing i 12,j:testing j 13,k:testing k 14,l:testing lNow we're good for a couple billion elements.