Parsing File with C# And Replace method

Question

I'm trying to parse a a bunch of file with Replace method(string) while is doing what I expect: I feels is not practical. for instance I will process 10K files but in the First 72 I found like 30 values that need to be replace And this is the rule :

My Goal :"

My goal is to replace all Instance of the ':' Dont follows this Rules :

1- the 2nd or 3rd Character foward is Not Another ':' 2-the 3rd or 2nd Chacarcter backward is Not Another ':'

All other should be Replaced

1- Any time that I found this character (:) and this character is not preceded by two char or three characters like :00: or :12A: I should replace it with an (*).

This is the method that I have so far.....

private static string cleanMesage(string str)
{
    string result = String.Empty;
    try
    {
        result = str.Replace("BNF:", "BNF*").Replace("B/O:", "B/O*").Replace("O/B:", "O/B*");
        result = result.Replace("Epsas:", "Epsas*").Replace("2017:", "2017*").Replace("BANK:", "BANK*");
        result = result.Replace("CDT:", "CDT*").Replace("ENT:", "").Replace("GB22:", "GB22*");
        result = result.Replace("A / C:", "A/C*").Replace("ORD:", "ORD*").Replace("A/C:", "A/C*");
        result = result.Replace("REF:", "REF*").Replace("ISIN:", "ISIN*").Replace("PAY:", "PAY*");
        result = result.Replace("DEPOSITO:", "DEPOSITO*").Replace("WITH:", "WITH*");
        result = result.Replace("Operaciones:", "Operaciones*").Replace("INST:", "INST*");
        result = result.Replace("DETAIL:", "DETAIL*").Replace("WITH:", "WITH*").Replace("BO:", "BO*");
        result = result.Replace("CUST:", "CUST*").Replace("ISIN:", "ISIN*").Replace("SEDL:", "SEDL*");
        result = result.Replace("Enero:", "Enero*").Replace("enero:", "Enero*");
        result = result.Replace("agosto:", "agosto*").Replace("febrero:", "febrero*");
        result = result.Replace("marzo:", "marzo*").Replace("abril:", "abril*");
        result = result.Replace("mayo:", "mayo*").Replace("junio:", "junio*").Replace("RE:", "RE:*");
        result = result.Replace("julio:", "julio*").Replace("septiembre:", "septiembre*");
        result = result.Replace("NIF:", "NIF*").Replace("INST:", "INST*").Replace("SHS:", "SHS*")
            .Replace("SK:", "");
        result = result.Replace("PARTY:", "PARTY*").Replace("SEDOL:", "SEDOL*").Replace("PD:", "PD*");
    }
    catch (Exception e)
    {

    }
    return result;
}

And this is some sample data :"

:13: <-- keep /ISIN/XS SVUNSK UXPORTKRUDIT ZX PZY DZTU:<- replace UX DZ
TU:<- replace02ZUG12 RZTU:<- replace W/H TZX RZTU:<- replace0.00000 SHZRUS PZID:<- replace
0.000000 IDDSIN:<- replace
:31:  <-- keep 1201000100CD05302,24NSUC20523531001//00520023531014
:13: <-- keep /ISIN/XS0153242003 SVUNSK UXPORTKRUDIT ZX PZY DZTU:<- replace00ZUG12 UX DZ
TU:02ZUG12 RZTU:0.30241 W/H TZX RZTU:<- replace0.00000 SHZRUS PZID:<- replace
0.000000 ISIN:XS0153242003
:31: <-- keep 1201000100DD121253,25S202IMSSMSZUX534C//S0322211DF4301
S F/O 0150001400
:13: <-- keep XNF:<- replace this

And you should show your input data without the keep suggestion as well the expected output — Steve
– Steve, Commented Oct 20, 2017 at 14:44
It's not practical, indeed. Each call to replace creates a new temporary string. If you have a large file, or have to process many files, this will cause a HUGE waste of memory and CPU. You should check regular expressions or parsers. For example, it looks like you are trying to modify tags. Those tags look like a bunch of letters followed by :. You can capture that with \w+:. That won't capture spaces or slashes though. In order to *replace something though, you need to catch the tag name, eg: (\w+):. This captures the tag. You can replace this with $1*. — Panagiotis Kanavos
– Panagiotis Kanavos, Commented Oct 20, 2017 at 14:47

Rocky · Accepted Answer · 2017-10-23 16:20:08Z

1

If your goal is to replace all instances of the ':' character where it is not followed by 2 or 3 other characters. You could indeed try the System.Text.RegularExpressions library. You could then simplify your cleanMessage function in the following way.

using System.Text.RegularExpressions;

function string cleanMessage(string str)
{
     string pattern = ":(\s)"; //This will be a ':' followed by a space
     Regex rgx = new Regex(pattern);
     string replaceResult = rgx.Replace(str,"*$1") //this will replace the pattern with a '*' followed by a space. 
     return replaceResult;
}

If your goal is to replace all instances of the ':' character where it is not followed by 2 or 3 other characters and the 2nd or 3rd character forward or backward is not another ':'. You could change your cleanMessage to the following instead.

using System.Text.RegularExpressions;

function string cleanMessage(string str)
{
     string pattern = "([^;]{2}.):(\s[^:]{2})"; 
     //This will be 2 characters that cannot be ':' followed by anything then a ':' followed by a space and 2 more characters that cannot by ':' 
     //For instance, "BNF: :F" would FAIL and not get replaced but "BNF: HH" would pass and become "BNF* HH"
     Regex rgx = new Regex(pattern);
     string replaceResult = rgx.Replace(str,"$1*$2") //this will replace the : with a * 
     return replaceResult;
}

More information on the System.Text.RegularExpressions library replace can be found at https://msdn.microsoft.com/en-us/library/xwewhkd1(v=vs.110).aspx

edited Oct 23, 2017 at 16:20

answered Oct 20, 2017 at 15:01

Rocky

1821 gold badge1 silver badge9 bronze badges

Sign up to request clarification or add additional context in comments.

3 Comments

Ivan S Over a year ago

My goal is to replace all Instance of the ':' Dont follows this Rules : 1- the 2nd or 3rd Character foward is Not Another ':' 2-the 3rd or 2nd Chacarcter backward is Not Another ':' All other should be Replaced

Rocky Over a year ago

I have edited my answer. Is it now closer to what you need? It should now only replace the : where it is no close to another one.

Ivan S Over a year ago

Awesome Thanks A lot

Simmetric · Accepted Answer · 2017-10-20 17:41:22Z

0

As @dymanoid mentioned, regular expressions are a way to handle this. By using the following you'd get what you want:

result = Regex.Replace(str, "([a-zA-Z0-9]{2,3})\:", "$1*");

However for large datasets this won't perform well. In that case I'd look at walking through str character by character using a for-loop. If the current character is not a colon, add it to the result string and to a temporary string. When the current character is a colon (:) and the temporary string has a length of 2 or 3, write an asterisk to the result and clear the temporary string. In this case you don't do any string replacement, you just select what to write to a new string.

See here for a speed comparison between string replacement and regex replacement.

edited Oct 20, 2017 at 17:41

answered Oct 20, 2017 at 14:51

Simmetric

1,6992 gold badges15 silver badges24 bronze badges

8 Comments

Chris Over a year ago

This doesn't look like it does what the OP wants... The first susbtitution in his current code is.Replace("BNF:", "BNF*"). As you can see it should only be replacing the : whereas yours will replace the preceding characters too...

Ivan S Over a year ago

Thanks a lot for the suggestion! this one will not work because is replacing this value ( keep -----> :31: <-- keep ) and I want to Replace this value - replace --->RZTU:<- replace, because on 2 or 3 position back dosnt have another : character... but it's a good starting point. so Thanks a lot!

Ivan S Over a year ago

@Chris it Does! because I only want to replace the (:) not the complete word (BNF:), my Result here is (BNF*) I that is exactly what im looking for...

Chris Over a year ago

@IvanS: I am confused again... It really looks like it doesn't... If I run var str = "BNF:"; var result = Regex.Replace(str, "[a-zA-Z0-9]{2,3}:", "*"); then result will equal just * and not BNF*. Is that actually correct? I don't think so from what you've said previously but your last comment seemed to indicate it is correct...

Ivan S Over a year ago

@Chris you are Right! I was Talking about the replace method, not the REGEX, this {result = Regex.Replace(str, "[a-zA-Z0-9]{2,3}\:", "*");} doesnt work it Replace the BNF also... I thought you was talking about the replace and not the regex, this regex wont work for what im looking for... but is a good staring point to learn about it.

|

Collectives™ on Stack Overflow

Parsing File with C# And Replace method

2 Answers 2

3 Comments

8 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

3 Comments

8 Comments

Your Answer

Sign up or log in

Post as a guest

Related