如何使用C＃解析JavaScript代码/数组

Question

I've got a web request to a javascript file. As a response I've JavaScript-Snippet which I'm trying to parse in C#.

The Snippet looks like this:

sDt[1647110]=['SVK U19 A','D43A71','Jupie Podlavice Badin(U19)','TJ Straza(U19)','','',' / '
,'','',114745,114746,1,'',0,0,0,1012,1,'','',''];sDt[1647108]=['SVK U19 A','D43A71','Kysucke Nove Mesto(U19)',
'MFK Lokomotiva Zvolen(U19)','','',' / ','','',114741,114742,1,'',0,0,0,1012,1,'','',''];
sDt[1647109]=['SVK U19 A', /* A lot of more of that kind followed by */ ;WLID[1623901]=1;
WLID[1623902]=1;WLID[1623903]=1;WLID[1637686]=1;
WLID[1637692]=1;WLID[1637687]=1;WLID[1637688]=1;WLID[1637685]= /* ending with */ 
var ORD = [1647110,1647108,1647109,1647133,1645669,1647122,1626152,1647251,1646643,
1647130,1646685,1 ... ];

Obviously this isn't pure JSON array. Now I wonder how to parse this most efficiently. First I started to do this per pedes meaning usig String.Split and so on. But this is slow and unfortunately not really stable.

While the Part behind each sDt[Idendifier]= is an Array which I could parse with Json.Net I also need the Idendifier . Everything else like WLID or var ORD I can ignore.

Does anyone has an idea how to do this efficiently?

Thanks in advance

Answer 1

You have to go through the whole request token by token if you don't have any other information. There is no other way around. Why don't you just send the JSON?

But to parse it I would do the following: Go through the whole request. If you come across a '[' make sure to check if you're not in a string. (For example by setting a flag when you stumble over a ' " ' and by unsetting it if you come to the next ' " '). If you're are not parsing a string right now, the following tokens are either the identifier or the content. You can easily check that. In case of a number, this is your identifier until you reach "]" (and given that you aren't parsing a string currently). In the other case it's the content which you can parse with Json.Net now, just remember where (the index) the first "[" and the following "]" is and you can generate a substring which you can then pass to Json.Net. If you come across a ";" and you are not in a string, make sure that you skip the WLID and ORD part.

The whole operation takes O(n * m) with n=Number of tokens and m=length of the longest content string.

If you do the parsing of the content yourself (and not letting Json.Net do that for you) you could narrow it down to O(n) of course.