使用Regex.Split创建字符串数组

Question

Alright, I'm warning you in advance, my understanding of Regular Expressions is extremely limited (I've tried my best to learn them over the years, but to be honest, I think they just frighten me.)

Let's say I have the following string:

string keyValues = "CustomerId=1||OrderId=12||UserId=a1dcd568-f129-419b-b51e-be2dbb67de0f"

This string represents key-value pairs, delimited by a user-defined string (in this case || ) (eg key1=value1||key2=value2 ). I am trying to extract the keys out of this string and store them in an array. That array would look like this:

{"CustomerId", "OrderId", "UserId"}

The best option I can think of is to use regular expressions (If someone has a better solution, please share). Here's what I'm trying to do:

string delimiter = "||";
string[] keys = Regex.Split(keyValues, "=.*" + delimiter);

I may be wrong, but the way I understand it, that regular expression is supposed to find a string that starts with = and ends with delimiter , with any number of any characters in between. Which would split the string at those positions, leaving me with the original keys, but instead, my keys array looks like this:

{"", "C", "u", "s", "t", "o", "m", "e", "r", "I", "d", "", "", ...}

As you can see, the =value|| part is stripped away. Can anyone tell me what I'm doing wrong?

EDIT

In my case, the delimiter || is a variable. I didn't mention this only because I thought I would be able to replace any references to || with delimiter . From the majority of the answers given, I now see that that is an important detail.

Answer 1

| has special meaning in regular expression ( patA|patB matches either patA or patB ). Escape | .

Using non-greedy match ( .*? ):

string delimiter = "||";
string[] keys = Regex.Split(keyValues, @"=.*?" + Regex.Escape(delimiter));

This will give you {"CustomerId", "OrderId", "UserId=a1dcd568-f129-419b-b51e-be2dbb67de0f"} .

Matches with lookahead assertion is more appropriate:

string delimiter = "||";
string keyValues = "CustomerId=1||OrderId=12||UserId=a1dcd568-f129-419b-b51e-be2dbb67de0f";
string pattern = @"(?<=^|" + Regex.Escape(delimiter) + @")\w+(?==)";
var keys = Regex.Matches(keyValues, pattern);

BTW, use verbatim string literals ( @"verbatim string literal" ) when express regular expression.

Demo

Answer 2

If you just care for the keys, why not try to use a match instead of a split using:

@"[^=|]+(?==)"

If the key can't contain an equal sign = or a vertical bar | , then the above expression will match one ore more characters that are not = or | which are followed by an equal sign = , thus matching the keys.

In C#:

var input = "CustomerId=1||OrderId=12||UserId=a1dcd568-f129-419b-b51e-be2dbb67de0f";
var results = Regex.Matches(input, @"[^=|]+(?==)");

Answer 3

An alternative is to do this without a regular expression, as the string operations are pretty basic:

string[] keys =
  keyValues.Split(new string[]{"||"}, StringSplitOptions.None)
  .Select(s => s.Substring(0, s.IndexOf('='))).ToArray();

Keep the regular expressions to the advanced string operations. :)

(When testing the performance of this solution compared to using a regular expression, this showed to be about 40 times faster.)

Answer 4

Split on @"=[^|]*(?:\\|\\||$)"
If you need more assurance, use @"=[^=|]*(?:\\|\\||$)"

Edited to consume end where no delimeter exists.
Try to just use no-blank elements if its in C#.