Incorrect sorting in Icelandic

[laddi]

I have a problem with text sorting using Icelandic characters. The problem is that character equivalents aren't really suitable since there is a strict sorting order for each character, the accented characters and then the special Icelandic characters (þ [thorn] and ð [eth]). The sorting is actually the same as specified on the sugar.js array sorting page which is as follows:

AÁ[ÀÂÃĄ]BC[ĆČÇ]D[Ď]ÐEÉ[ÈĚÊË]ĘFG[Ğ]H[ı]IÍ[ÌİÎÏ]JKL[Ł]MN[ŃŇÑ]OÓ[ÒÔ]PQR[Ř]S[ŚŠŞ]T[Ť]UÚ[ÙŮÛÜ]VWXYÝZ[ŹŻŽ]ÞÆ[ŒØÕÅÄ]Ö (characters in brackets not in Icelandic alphabet)

But using the sortLocaleCompare sorts á/Á, for example, as an a/A which is wrong because a > á. Also ö/Ö is mixed in with the O's but is actually the last letter in the alphabet. And þ/Þ is last behind both æ/Æ and ö/Ö even though both of them should be behind þ/Þ.

Any ideas on a possible fix for this or, better yet, use the correct sorting as specified in the sugar.js documentation? Or is there possibly a character to number (place in alphabet) conversion I could use?

[Mottie]

Hmm, well I guess I'll need to dig into sugar.js' sort function because this plugin doesn't sort characters in that manner.

It might not be a simple, quick fix either. So, I'll apologize in advance for the inconvenience and hopefully will have a viable solution for you soon.

[laddi]

Thanks for a speedy reply!

I actually kind of knew that already since I've been trying to fix this for a while now. This is the drawback of working in a language that only around 330.000 people in the world can speak and/or understand... ;)

[Mottie]

Whew... my darn OCD. Anyway, I pulled out the relevant code from sugar.js and made a sort function (sortBy). So all you need to do is include the sort code below, and set the textSorter option to sortBy. :)

You do not need to include sugar.js. Here is a demo.

$(function(){

    /* sortBy functions & helpers extracted and slightly modified from
    * Sugar Library v1.3.7
    * Freely distributable and licensed under the MIT-style license.
    * Copyright (c) 2012 Andrew Plummer
    * http://sugarjs.com/arrays#sorting
    */
    var array = {
        AlphanumericSortIgnoreCase  : true,
        AlphanumericSortEquivalents : {}
    },

    // order = 'AÁÀÂÃĄBCĆČÇDĎÐEÉÈĚÊËĘFGĞHıIÍÌİÎÏJKLŁMNŃŇÑOÓÒÔPQRŘSŚŠŞTŤUÚÙŮÛÜVWXYÝZŹŻŽÞÆŒØÕÅÄÖ',
    // equiv = 'AÁÀÂÃ,CÇ,EÉÈÊË,IÍÌİÎÏ,OÓÒÔ,Sß,UÚÙÛÜ',
    // modified order to match Icelandic sorting - see https://github.com/Mottie/tablesorter/issues/212
    order = 'AÁBCDÐEÉĘFGHIÍJKLMNOÓPQRSTUÚVWXYÝZÞÆÖ',
    equiv = '',

    sortBy = function(a,b){
        return typeof a === "string" && typeof b === "string" ? collateStrings(a,b) : a < b ? -1 : a > b ? 1 : 0;
    },

    // Alphanumeric collation helpers
    collateStrings = function(a, b) {
        var aValue, bValue, aChar, bChar, aEquiv, bEquiv, index = 0, tiebreaker = 0;
        a = getCollationReadyString(a);
        b = getCollationReadyString(b);
        do {
            aChar  = getCollationCharacter(a, index);
            bChar  = getCollationCharacter(b, index);
            aValue = getCollationValue(aChar);
            bValue = getCollationValue(bChar);
            if (aValue === -1 || bValue === -1) {
                aValue = a.charCodeAt(index) || null;
                bValue = b.charCodeAt(index) || null;
            }
            aEquiv = aChar !== a.charAt(index);
            bEquiv = bChar !== b.charAt(index);
            if (aEquiv !== bEquiv && tiebreaker === 0) {
                tiebreaker = aEquiv - bEquiv;
            }
            index += 1;
        } while (aValue != null && bValue != null && aValue === bValue);
        if (aValue === bValue) return tiebreaker;
        return aValue < bValue ? -1 : 1;
    },

    getCollationReadyString = function(str) {
        if (array.AlphanumericSortIgnoreCase) {
            str = str.toLowerCase();
        }
        return str.replace(array.AlphanumericSortIgnore, '');
    },

    getCollationCharacter = function(str, index) {
        var chr = str.charAt(index), eq = array.AlphanumericSortEquivalents || {};
        return eq[chr] || chr;
    },

    getCollationValue = function(chr){
        var order = array.AlphanumericSortOrder;
        return chr ? order.indexOf(chr) : null;
    },

    equivalents = {};

    array.AlphanumericSortOrder = $.map(order.split(''), function(str) {
        return str + str.toLowerCase();
    }).join('');

    $.each(equiv.split(','), function(i,set) {
        var equivalent = set.charAt(0);
        $.each(set.slice(1).split(''), function(i,chr) {
            equivalents[chr] = equivalent;
            equivalents[chr.toLowerCase()] = equivalent.toLowerCase();
        });
    });

    $('table.tablesorter').tablesorter({
        widgets    : ['zebra'],
        textSorter : sortBy
    });
});

[laddi]

Thanks a million! :D

I just added a global override of the textSorter default and now it's working perfectly for me in all views. Thanks again and keep up the good work on an excellent plugin! :D

[Mottie]

Glad I could help! :)

[Mottie]

Just to update you, Sugar has since exposed the sort functions so there is no longer a need to extract out specific code from that library. All you need to get a proper Icelandic alphabetical sort is to do the following (demo):

$(function () {

    Array.AlphanumericSortOrder = 'AaÁáBbCcDdÐðEeÉéĘęFfGgHhIiÍíJjKkLlMmNnOoÓóPpQqRrSsTtUuÚúVvWwXxYyÝýZzÞþÆæÖö';
    Array.AlphanumericSortIgnoreCase = true;
    // see https://github.com/andrewplummer/Sugar/issues/382#issuecomment-41526957
    Array.AlphanumericSortEquivalents = {};

    $('table.tablesorter').tablesorter({
        widgets: ['zebra'],
        textSorter: {
            0 : Array.AlphanumericSort
        }
    });

});

[laddi]

Cool, thanks! :D