Fast way for extracting tokens from a string.

  • Thread starter Thread starter Jensen bredal
  • Start date Start date
J

Jensen bredal

Hello,
I have a string formated in the following way:

s = 1;32;100;32;09;.........;09;76;

I need to extract the numbers separated by the seicolon.
The list can contain several thousands of items and the code is
time critical.

How can i best extract them in C#.

Many Thanks in advance

JB
 
Jensen,

I would just use the Split method on the string itself, passing a
semicolon. It's probably going to give you the fastest performance.

However, I don't know that you should use a string at all. If the
string is rather large, you should be parsing it apart as you are retrieving
the information. For example, if the string was in a file, or being read
over a stream, I would parse it out as I read the characters from the
stream, not once the string was constructed.

If you got it from someplace else, like a database field (where it is in
string format already) and you can't do anything about it being in a string,
use the Split method.

Hope this helps.
 
Nicholas Paldino said:
I would just use the Split method on the string itself, passing a
semicolon. It's probably going to give you the fastest performance.
Click to expand...

That would not be my expectation. Split is going to create "several
thousands of" little strings, greatly increasing memory pressure. This
is a classic case where enumerating through Regex.Matches should be
faster than creating an array of strings. Plus a regex could ignore
white space, check the tokens are all digits, &c.
 
Jensen said:
could you provide some sample code?
Click to expand...

<semicode> Your sample string

s = 1;32;100;32;09;.........;09;76;

contains a stream of digits followed by semicolons. The regex

(\d+);

will match a stream of digits followed by semicolons, capturing the
digits. This capture is only a matter of calculating string length and
start offset - no substring operations are done until you read a Value
property. So, do a foreach on Regex.Matches() of your data and the
@"(\d+);" regex pattern. Read the 2nd group in each match - that's got
the captured digits, without the semicolon.

</semicode>
 
Back
Top