ViNull, Off the Record - All Comments

The ASP.NET MVC Definition « Payed Coder

The ASP.NET MVC Definition « Payed Coder — Sat, 13 Dec 2008 15:45:22 GMT

Pingback from The ASP.NET MVC Definition « Payed Coder

re: SilverLight Interop with Flash/Flex (flashlight?)

webdesign — Wed, 26 Nov 2008 15:14:14 GMT

Nice flash code! Thanks

re: Finding the difference between two Arrays, or un-LINQ-ing your code

Kevin H — Wed, 08 Oct 2008 00:51:39 GMT

Well, the DateTime timing isn't very good (15ms resolution, plus the GC might run, etc). I've re-timed it a few times, and it seems that your method is slightly better than the method in the post, is slightly better than Except. However, they are all orders of magnitude faster than the LINQ .contains version, and 500k entries take less than a couple hundred ms, which is significantly less time than it would take to actually load that amount of data from the file system or db.

re: Finding the difference between two Arrays, or un-LINQ-ing your code

Bill — Sat, 04 Oct 2008 21:20:39 GMT

The Except extension method is what you are looking for (though your solution is almost twice as fast because using the extension method you would be sorting the lists twice, contrary to the incorrect results suggested by Keven). My results on the comparison methods above are:

ExceptMethod: 31.25ms

MergeSortMethod: 15.625ms

though I changed the instantiation for loops:

static void Main(string[] args) {

Random r = new Random(1);

List<int> oldNumbers = Enumerable.Range(1, NumberOfSamples).Select(x => r.Next(NumberOfSamples*2)).Distinct().ToList();

List<int> newNumbers = Enumerable.Range(1, NumberOfSamples).Select(x => r.Next(NumberOfSamples * 2)).Distinct().ToList();

A slight improvement on your code would be (a couple of saved instructions per loop, we are talking microseconds of improvement here):

var removed = new List<int>();

var added = new List<int>();

newNumbers.Sort();

oldNumbers.Sort();

int iE = 0, iI = 0, oldCount = oldNumbers.Count, newCount = newNumbers.Count;

while (iE < oldCount && iI < newCount) {

var diff = oldNumbers[iE].CompareTo(newNumbers[iI]);

if (diff == 0) {

iE++;

iI++;

} else if (diff > 0) {

removed.Add(newNumbers[iI]);

iI++;

} else {

added.Add(oldNumbers[iE]);

iE++;

}

if (iE >= oldCount) {

while (iI < newCount) {

removed.Add(newNumbers[iI]);

iI++;

}

} else if (iI >= newCount) {

while (iE < oldCount) {

added.Add(oldNumbers[iE]);

iE++;

}

Dew Drop - October 4, 2008 | Alvin Ashcraft's Morning Dew

Dew Drop - October 4, 2008 | Alvin Ashcraft's Morning Dew — Sat, 04 Oct 2008 14:40:35 GMT

Pingback from Dew Drop - October 4, 2008 | Alvin Ashcraft's Morning Dew

re: Finding the difference between two Arrays, or un-LINQ-ing your code

charles — Sat, 04 Oct 2008 09:54:18 GMT

I have to try this one.

re: Finding the difference between two Arrays, or un-LINQ-ing your code

smaclell — Fri, 03 Oct 2008 23:59:53 GMT

Since what you are doing appears to be filtering each list based on where they overlap you might want to check out the Intersects and Except extension methods.

rsanidad.wordpress.com/.../linq-except-and-intersect

Good luck.

re: Finding the difference between two Arrays, or un-LINQ-ing your code

Kevin H — Fri, 03 Oct 2008 23:55:14 GMT

There's always the .Except() method, which does what you want, and seems to be a little faster even than your mergesort (i'm guessing because it *is* a mergesort written unmanaged)

Adding/removing 1000 out of

Linq w/ contains: 17734.375ms

Except: 15.625ms

MergeSort: 46.875ms

source code follows:

class Program

{

const int NumberOfSamples = 50000;

static void Main(string[] args)

{

Random r = new Random();

List<int> oldNumbers = new List<int>();

List<int> newNumbers = new List<int>();

for (int i = 0; i < NumberOfSamples; i++)

oldNumbers.Add(r.Next());

newNumbers.AddRange(oldNumbers);

for (int i = 0; i < 1000; i++)

newNumbers.RemoveAt(r.Next(newNumbers.Count));

for (int i = 0; i < 1000; i++)

newNumbers.Add(r.Next());

DateTime start;

GC.Collect();

start = DateTime.Now;

ContainsMethod(oldNumbers, newNumbers);

Console.WriteLine("{0}ms", (DateTime.Now - start).TotalMilliseconds);

GC.Collect();

start = DateTime.Now;

ExceptMethod(oldNumbers, newNumbers);

Console.WriteLine("{0}ms", (DateTime.Now - start).TotalMilliseconds);

GC.Collect();

start = DateTime.Now;

MergeSortMethod(oldNumbers, newNumbers);

Console.WriteLine("{0}ms", (DateTime.Now - start).TotalMilliseconds);

Console.ReadKey();

}

static void ExceptMethod(List<int> oldNumbers, List<int> newNumbers)

{

List<int> removed = oldNumbers.Except(newNumbers).ToList();

List<int> added = newNumbers.Except(oldNumbers).ToList();

Console.WriteLine("Removed: {0}, Added: {1}", removed.Count, added.Count);

}

static void ContainsMethod(List<int> oldNumbers, List<int> newNumbers)

{

List<int> removed = (from f in oldNumbers

where !newNumbers.Contains(f)

select f).ToList();

List<int> added = (from f in newNumbers

where !oldNumbers.Contains(f)

select f).ToList();

Console.WriteLine("Removed: {0}, Added: {1}", removed.Count, added.Count);

}

static void MergeSortMethod(List<int> oldNumbers, List<int> newNumbers)

{

List<int> removed = new List<int>();

List<int> added = new List<int>();

newNumbers.Sort();

oldNumbers.Sort();

for (int iE = 0, iI = 0; iE < oldNumbers.Count || iI < newNumbers.Count; ) {

if (iE >= oldNumbers.Count) {

removed.Add(newNumbers[iI]);

iI++;

}

else if (iI >= newNumbers.Count) {

added.Add(oldNumbers[iE]);

iE++;

}

else {

Int32 diff = oldNumbers[iE].CompareTo(newNumbers[iI]);

if (diff == 0) {

iE++;

iI++;

}

else if (diff > 0) {

removed.Add(newNumbers[iI]);

iI++;

}

else if (diff < 0) {

added.Add(oldNumbers[iE]);

iE++;

}

Console.WriteLine("Removed: {0}, Added: {1}", removed.Count, added.Count);

}

re: Finding the difference between two Arrays, or un-LINQ-ing your code

Carlos Beppler — Fri, 03 Oct 2008 23:36:27 GMT

Try to use System.Collections.Generic.HashSet<T> instead of lists.

To get an HashSet you can change the fist code to:

HashSet<String> toAdd = new HashSet<string>(

from f in existingFiles

where !indexedFiles.Contains(f)

select f).AsEnumerable() );

re: Finding the difference between two Arrays, or un-LINQ-ing your code

LukeB — Fri, 03 Oct 2008 22:42:28 GMT

Iesi.Collections.ISet<> is your friend.

Websites tagged "orm" on Postsaver

Websites tagged "orm" on Postsaver — Fri, 19 Sep 2008 08:17:29 GMT

Pingback from Websites tagged "orm" on Postsaver

re: Using LINQ to generate HTML

Yuvi — Mon, 04 Aug 2008 18:26:19 GMT

VB9's XML Literals ftw!

re: Using LINQ to generate HTML

KevinI — Mon, 04 Aug 2008 17:25:06 GMT

You should check out VB.Net XML Literals, it uses the LINQ with type strict syntax to take the exact code you made above and basically make it look like 'pure xml' (but type safe).

Dew Drop - August 4, 2008 | Alvin Ashcraft's Morning Dew

Dew Drop - August 4, 2008 | Alvin Ashcraft's Morning Dew — Mon, 04 Aug 2008 13:28:58 GMT

Pingback from Dew Drop - August 4, 2008 | Alvin Ashcraft's Morning Dew

re: Using LINQ to generate HTML

Nicholas Piasecki — Sun, 03 Aug 2008 19:48:28 GMT

I have no problem mixing code with markup, particularly little if/else checks on template pages, as long as that code is presentation code. This usage of LINQ isn't really creating any separation, just obfuscating it. After all, at the end of the day, *some* code is going to spit out the string ... which method will be easier for a maintenance programmer to modify six months down the road is a stylistic preference that is open to debate. Just my two cents.