GeoServer CQL Filter doesnt work with Unicode

Question

I'm using CQL filter to extract certain data from my GeoServer layer.

var source = new ol.source.Vector({
        loader: function (extent, resolution, projection) {
            var url = "http://localhost/geoserver/visualization/ows?service=WFS"
                + "&version=1.1.0&request=GetFeature"
                + "&typeName=visualization:".concat(inputTable)
                + '&outputFormat=text/javascript'
                + "&format_options=callback:loadFeatures"
                + '&srsname=EPSG:3857'
                + filter;
            $.ajax({ url: url, dataType: 'jsonp', jsonp: false });
            console.log(filter);
            console.log(url);
        }
    });

Notice the filter string as the final parameter at url. If I normally input the feature name using alphabet text, it will return the result. But if I use other than alphabet (UTF-8, EUC-KR) it won't return anything.

&cql_filter=branchCode='N01821'  -> works
&cql_filter=방향='S'             -> korean word, doesn't work

If I print the URL, the Korean strings give me this error

<ows:ExceptionText>Could not parse CQL filter list. Lexical error at line 1, column 1. Encountered: "\ubc29" (48169), after : "" Parsing : 방향='S'.</ows:ExceptionText>

Anyone knows the reason why and the solution?

Answer 1

Because the grammar of CQL defines an identifier as:

    < IDENTIFIER: (<LETTER> (<LETTER>|<DIGIT>)*) | <DOUBLE_QUOTE> (<ANY>)+ <DOUBLE_QUOTE> > |

    < #ANY: ~[" ","\""] > |
    < #LETTER: [ "a"-"z", "A"-"Z" , "_"] > |

However check the second form of the identifier, <DOUBLE_QUOTE> (<ANY>)+ <DOUBLE_QUOTE>: you should be able to get your attribute names work by double quoting them, e.g. "방향"='S'