Enhance the way of linear search

Question

Enhance the way of linear search

Jonathan 85

Hi,

To C# linear search below

User's image

any advice to apply the following way on the above?

Jonathan 85 Reputation points

2026-06-01T02:17:21.81+00:00

Sorry to all, the screen below was missing (from the above post). How to apply the way below by adjusting the original code?
Bruce (SqlWork.com) 84,071 Reputation points

2026-06-01T14:25:21.7566667+00:00

It’s not clear how this code is related to the first code.

As the file read is the most expensive part, adding parallel processing to scan one line of text scan and blocking on the add to collection, the routine is probably slower than using a regular collection and a no parallel processing. That is the overhead of creating and managing threads is higher than a string scan. Also the blocking for concurrency is slower than a standard collection.

4 answers

Your answer

Jonathan 85 Reputation points

2026-06-01T02:17:21.81+00:00

Sorry to all, the screen below was missing (from the above post). How to apply the way below by adjusting the original code?
Bruce (SqlWork.com) 84,071 Reputation points

2026-06-01T14:25:21.7566667+00:00

It’s not clear how this code is related to the first code.

As the file read is the most expensive part, adding parallel processing to scan one line of text scan and blocking on the add to collection, the routine is probably slower than using a regular collection and a no parallel processing. That is the overhead of creating and managing threads is higher than a string scan. Also the blocking for concurrency is slower than a standard collection.

Answer 1

Nancy Vo (WICLOUD CORPORATION) 4,845 Microsoft External Staff Moderator

Hello @Jonathan ,

Thanks for your question.

Instead of applying SearchValues + Parallel.ForEach to your existing List, I would suggest a simpler and more effective approach - replace List<string> with HashSet<string>.

You can refer to my example code:

var ipSet = new HashSet<string>(
    File.Exists(@"c:/cmp/list11.txt")
        ? File.ReadLines(@"c:/cmp/list11.txt").Select(l => l.Trim())
        : Enumerable.Empty<string>(),
    StringComparer.OrdinalIgnoreCase
);

public static bool AddIfNotExists(string ip, HashSet<string> ipSet)
{
    ip = ip.Trim();

    if (ipSet.Add(ip))
    {
        File.AppendAllText(@"c:/cmp/list11.txt", ip + Environment.NewLine);
        return true;
    }

    return false;
}

I hope this addresses your question. If this response was helpful, please consider following the guidance to provide feedback.

Jonathan 85 Reputation points

2026-06-01T08:07:27.1166667+00:00

Thanks a lot to Nancy for your code. From IpSet, which boolean or value should I make use of, to decide it for the original line below?

ind0 = list0.FindIndex(item => item == ip);
Nancy Vo (WICLOUD CORPORATION) 4,845 Reputation points Microsoft External Staff Moderator

2026-06-01T08:23:37.7166667+00:00
Hi @Jonathan ,

Please replace the following code:

ind0 = list0.FindIndex(item => item == ip); if (ind0 < 0) { list0.Add(ip); // ... write file return true; }

with:

if (ipSet.Add(ip)) { // ... write file return true; }
Jonathan 85 Reputation points

2026-06-01T08:28:18.24+00:00

When did you refer to AddIfNotExists?
Nancy Vo (WICLOUD CORPORATION) 4,845 Reputation points Microsoft External Staff Moderator

2026-06-01T08:33:33.64+00:00

AddIfNotExists is the same method you already have, just rewritten. Therefore, you only need to replace the body of your existing method with my example code - there’s no need to create a new method named AddIfNotExists.
Jonathan 85 Reputation points

2026-06-01T09:48:17.5866667+00:00

Sorry, when did you see AddIfNotExists event was existing on my original code?
Jonathan 85 Reputation points

2026-06-02T03:58:31.3566667+00:00

Sorry, how to adjust it below?

Nancy Vo (WICLOUD CORPORATION) 4,845 Microsoft External Staff Moderator

Hi @Jonathan ,

Please refer to my example code and let me know the results. I’d be happy to investigate further if needed.

static bool ShouldBlockIP(string ip)
{
    if (string.IsNullOrEmpty(ip)) return false;

    string[] safePrefixes = { "113.", "223.", "127.", "103.15", "182.173", "40.107", "142.251", "142.250" };

    var ipSet = new HashSet<string>(
        File.Exists(@"c:/cmp/list11.txt")
            ? File.ReadLines(@"c:/cmp/list11.txt").Select(l => l.Trim())
            : Enumerable.Empty<string>(),
        StringComparer.OrdinalIgnoreCase
    );

    if (ipSet.Add(ip))
    {
        File.AppendAllText(@"c:/cmp/list11.txt", ip + Environment.NewLine);
        return true;
    }

    return false;
}

Jonathan 85 Reputation points

2026-06-02T09:21:54.22+00:00

Many thanks to your nice help! Is it still a little bit of issue on this since I got repeated IP like

let me know if you want to see more details on my project.
Nancy Vo (WICLOUD CORPORATION) 4,845 Reputation points Microsoft External Staff Moderator

2026-06-02T09:31:48.27+00:00

Thanks for sharing. Could you help me better understand the purpose of ShouldBlockIP? I just want to make sure I understand correctly - please clarify your intended use case so I can investigate further.
Jonathan 85 Reputation points

2026-06-02T09:43:47.4033333+00:00

List11 file is having all Unique IPs ever blocked. I saw the file is having this

unexpectedly at the end of the file. It was having the issue to write to the file correctly!
Jonathan 85 Reputation points

2026-06-04T02:04:11.9466667+00:00

Nancy, Do you need more details for this? There was the issue to write one IP properly into the text file for your way. Could you see this? Or can anyone help?

Bruce (SqlWork.com) 84,071

as the question was about performance, the hash set is lowering it. as the ip test causes a file I/O, and the hash set is only used once it is slower than just a compare.. worse, you always read the whole file, even if there is a match in first row. try:

static bool ShouldBlockIP(string ip)
{
    const fileName = @"c:\cmp\list11.txt"
    ip = ip?.Trim()

    if (string.IsNullOrEmpty(ip)) return false;

    string[] safePrefixes = { "113.", "223.", "127.", "103.15", "182.173", "40.107", "142.251", "142.250" };

    // test if ip in file contents
    // as ip's are numbers, no need for case insensitive compare
    var found = File.Exists(fileName)
            ? File.ReadLines(fileName).FindIndex(l => ip == l.Trim()) >= 0
            : false;

    // if ip not found, append to text file with line ending
    if (!found)
    {
        File.AppendAllText(fileName, ip + Environment.NewLine);
    }

    return found;
}

Answer 2

Bruce (SqlWork.com) 84,071

The file I/O is orders of magnitude slower than the index lookup, you will never notice the difference. If the lookup is in a loop, you want to extract the file access to outside the loop. That is build a new list, the write the list once. also you don’t show how lookups are done. Converting the list to a dictionary or hash requires iteration over the list, so if there is only one lookup, the conversion is more expensive. Also if the list is small, find index is not expensive, and may be quicker.

0 comments

Answer 3

Viorel 126.9K

The solution depends on other details. For example, if the list is loaded from file, then:

string path = @"c:\cmp\list11.txt";

ip = ip.Trim( );

if( !File.Exists( path ) || !File.ReadLines( path ).Any( a => a == ip ) )
{
    File.AppendAllText( path, ip + Environment.NewLine );

    return true;
}

0 comments

Answer 4

List<T>.FindIndex performs a linear search (O(n)) on every lookup. For frequent lookups or large collections, replace the list-based search with a data structure that provides O(1) lookup.

A typical enhancement is to use HashSet<string> or Dictionary<string, TValue> instead of repeatedly calling FindIndex on List<string>:

// Build a HashSet once from the existing list
var ipSet = new HashSet<string>(list0);  // case-sensitive; use custom comparer if needed

// When checking/adding a new IP
if (!ipSet.Contains(ip))
{
    ipSet.Add(ip);
}

// When persisting back to file, enumerate the set
File.Delete(@"c:/cmp/list11.txt");

using (var sw = new StreamWriter(@"c:/cmp/list11.txt", append: false))
{
    foreach (var item in ipSet)
    {
        sw.WriteLine(item.Trim());
    }
}

Key improvements compared to the original linear search with FindIndex:

HashSet<string>.Contains is O(1) on average, instead of O(n) for List<T>.FindIndex.
The output file is opened once and written in a single pass, instead of opening a new StreamWriter inside the loop.
The collection (ipSet) naturally enforces uniqueness, so duplicates are avoided without extra checks.

If the order of elements must be preserved, maintain both a List<string> for ordering and a HashSet<string> for fast membership checks:

var ipList = new List<string>();
var ipSet  = new HashSet<string>();

void AddIp(string ip)
{
    if (ipSet.Add(ip))
    {
        ipList.Add(ip); // only added when new
    }
}

This pattern keeps lookups efficient while preserving insertion order for file output.

References:

Share via

Enhance the way of linear search

4 answers

Your answer