Weird behavior of jobname

jobname produces characters of category code 12, while the replacement text of SK contains characters of category code 11.

There are several ways to cope with the problem. If you use pdftex that supports e-TeX extensions

edefSK{detokenize{SK}}

defparse"#1 (#2)"{deffilelanguage{#2}}

expandafterparsejobname

ifxSKfilelanguage equalelse distinctfi

bye

You could also exploit pdfstrcmp:

defparse"#1 (#2)"{deffilelanguage{#2}}

expandafterparsejobname

ifnumpdfstrcmp{SK}{filelanguage}=0 equalelse distinctfi

bye

because pdfstrcmp does string comparison independent of category codes (and expands macros in its arguments).

In any case, you shouldn't do deflanguage, because language is a TeX primitive.

A more flexible solution with expl3.

input expl3-generic

ExplSyntaxOn
str_new:N l_bak_file_language_str
str_set_eq:NN l_bak_file_language_str c_sys_jobname_str
regex_replace_once:nnN { .*? ((.*)) .* } { 1 } l_bak_file_language_str

% now the string variable contains the string in parentheses
% extracted from the jobname

cs_new:Npn checklanguage
 {
  str_case:VnF { l_bak_file_language_str }
   {
    {SK}{Language~is~SK}
    {AB}{Language~is~AB}
    {XYZ}{Language~is~XYZ}
   }
   {Undefined~language}
 }
ExplSyntaxOff

checklanguage

bye

This prints “Language is SK”, but if I use a different jobname, say XYZ(X), I get “Undefined language”.

The command checklanguage is fully expandable and so works in edef. The token to execute for each string are up to you and your intended application. Note that you don't need to care about the quotes added if there is a space in the name, because the first lines just extract what's between the (first set of) parentheses.