Comments on: Collection of VBScript string functions for classic ASP http://www.webdevbros.net/2006/11/20/collection-of-vbscript-string-functions-for-classic-asp/ hot talk about web development Tue, 10 May 2011 13:11:20 +0200 hourly 1 http://wordpress.org/?v=3.0.1 By: English Classics http://www.webdevbros.net/2006/11/20/collection-of-vbscript-string-functions-for-classic-asp/#comment-9607 English Classics Tue, 23 Dec 2008 03:31:44 +0000 http://www.webdevbros.net/2006/11/20/collection-of-vbscript-string-functions-for-classic-asp/#comment-9607 Very neat, indeed. Both of my websites are in classic ASP so some of these may definitely come in handy. Very neat, indeed. Both of my websites are in classic ASP so some of these may definitely come in handy.

]]> By: Michal http://www.webdevbros.net/2006/11/20/collection-of-vbscript-string-functions-for-classic-asp/#comment-8823 Michal Fri, 11 Jul 2008 15:54:38 +0000 http://www.webdevbros.net/2006/11/20/collection-of-vbscript-string-functions-for-classic-asp/#comment-8823 thanks craig for your samples and your ideas ... this made me playing around today for some time :) yeah the regex object creation takes the most time .. but i discovered that the main problem with your approach is that its execution lasts longer when the placeholders are used more than once in a string. example: [asp] format("{0}{1}{0}{0}", array(1, 2)) [/asp] first i have tried the approach i mentioned to you .. which resulted in this: [asp] public function format3(byVal str, byVal args) if not isArray(args) then args = array(args) set rx = new RegExp for i = 0 to uBound(args) rx.pattern = "([^_]|^)\{" & i & "\}([^_]|$)" rx.global = true str = rx.replace(str, "$1_" & args(i) & "_$2") next format3 = replace(str, "_", "") end function [/asp] problem was that it looks shorter but takes even longer :))) however this approach led me to my final solution which seems the best i guess.. its also almost the same speed as the original one. [asp] function format(byVal str, byVal values) escapeChar = asc(24) & asc(27) if instr(str, escapeChar) > 0 then response.write("unsupported char found in input string") response.end() end if if not isArray(values) then values = array(values) for i = 0 to ubound(values) val = cStr(values(i)) if instr(val, escapeChar) > 0 then response.write("unsupported char found in argument") response.end() end if val = replace(val, "{", escapeChar) str = replace(str, "{" & i & "}", val) next format = replace(str, escapeChar, "{") end function [/asp] the only little odd on this is that in theory the string could contain a sequence of chars which is used internally for escaping. this results in an error then .. but the chance is very very low .. take a look at the code. Thanks for your inspiration craig and your time! the updated version can be found in the ajaxed SVN as this file here is not maintained. btw the code i ran for my benchmarks is the following: [asp] values = array( _ "{0}{1}{2}{3}{4}{5}{6}{7}{8}{0}{1}{2}{3}{4}{5}{6}{7}{8}", _ "asdasdasda{0}asdasd{1}asdasda{2}asdasda{3}adsad{4}adsad{5}adsad{6}asdad{7}asdad{8}asdasdda", _ "{0}{1}{2}{3}asdasd{4}{5}{7}{8}adasda{0}{1}{2}{3}asdasd{4}{5}{7}{8}" _ ) s = timer() for each v in values for i = 0 to 1000 x = str.format(v, array("a", "b", "c", "d", "e", "f", "g", "h", "j")) next next str.write("new: " & timer() - s) [/asp] last but not least: the new function also supports a non-array as argument as well. This is useful if you only need one argument to replace. thanks craig for your samples and your ideas ... this made me playing around today for some time :)
yeah the regex object creation takes the most time .. but i discovered that the main problem with your approach is that its execution lasts longer when the placeholders are used more than once in a string. example:

PLAIN TEXT
ASP:
  1. format("{0}{1}{0}{0}", array(1, 2))

first i have tried the approach i mentioned to you .. which resulted in this:

PLAIN TEXT
ASP:
  1. public function format3(byVal str, byVal args)
  2.     if not isArray(args) then args = array(args)
  3.     set rx = new RegExp
  4.     for i = 0 to uBound(args)
  5.         rx.pattern = "([^_]|^)\{" & i & "\}([^_]|$)"
  6.         rx.global = true
  7.         str = rx.replace(str, "$1_" & args(i) & "_$2")
  8.     next
  9.     format3 = replace(str, "_", "")
  10. end function

problem was that it looks shorter but takes even longer :))) however this approach led me to my final solution which seems the best i guess.. its also almost the same speed as the original one.

PLAIN TEXT
ASP:
  1. function format(byVal str, byVal values)
  2.     escapeChar = asc(24) & asc(27)
  3.     if instr(str, escapeChar)> 0 then
  4.         response.write("unsupported char found in input string")
  5.         response.end()
  6.     end if
  7.     if not isArray(values) then values = array(values)
  8.     for i = 0 to ubound(values)
  9.         val = cStr(values(i))
  10.         if instr(val, escapeChar)> 0 then
  11.             response.write("unsupported char found in argument")
  12.             response.end()
  13.         end if
  14.         val = replace(val, "{", escapeChar)
  15.         str = replace(str, "{" & i & "}", val)
  16.     next
  17.     format = replace(str, escapeChar, "{")
  18. end function

the only little odd on this is that in theory the string could contain a sequence of chars which is used internally for escaping. this results in an error then .. but the chance is very very low .. take a look at the code.
Thanks for your inspiration craig and your time! the updated version can be found in the ajaxed SVN as this file here is not maintained.

btw the code i ran for my benchmarks is the following:

PLAIN TEXT
ASP:
  1. values = array( _
  2.         "{0}{1}{2}{3}{4}{5}{6}{7}{8}{0}{1}{2}{3}{4}{5}{6}{7}{8}", _
  3.         "asdasdasda{0}asdasd{1}asdasda{2}asdasda{3}adsad{4}adsad{5}adsad{6}asdad{7}asdad{8}asdasdda", _
  4.         "{0}{1}{2}{3}asdasd{4}{5}{7}{8}adasda{0}{1}{2}{3}asdasd{4}{5}{7}{8}" _
  5.     )
  6.     s = timer()
  7.     for each v in values
  8.         for i = 0 to 1000
  9.             x = str.format(v, array("a", "b", "c", "d", "e", "f", "g", "h", "j"))
  10.         next
  11.     next
  12.     str.write("new: " & timer() - s)

last but not least: the new function also supports a non-array as argument as well. This is useful if you only need one argument to replace.

]]> By: Craig http://www.webdevbros.net/2006/11/20/collection-of-vbscript-string-functions-for-classic-asp/#comment-8803 Craig Thu, 10 Jul 2008 22:45:08 +0000 http://www.webdevbros.net/2006/11/20/collection-of-vbscript-string-functions-for-classic-asp/#comment-8803 Eh, forget about the previous data (lies, damn lies, and statistics). I had a bug in the performance code where the test string was modified making the tests run faster. Here is the new code: http://pastebin.com/m3a71221e In hindsite, the tests only showed that the methods both perform linearly on the same length of args array and same number of replacements. I've varied the number of the replacements in the new code. I found that both methods performed linearly based on the number of replacements made. The average time (on my development box) per a replacement was 0.0015 ms for the original method and 0.0362 ms for the new method. Since this it was O(n) and the difference is a fraction of a millisecond, I've decided to run with it for our code base ;) Eh, forget about the previous data (lies, damn lies, and statistics). I had a bug in the performance code where the test string was modified making the tests run faster. Here is the new code: http://pastebin.com/m3a71221e

In hindsite, the tests only showed that the methods both perform linearly on the same length of args array and same number of replacements. I've varied the number of the replacements in the new code.

I found that both methods performed linearly based on the number of replacements made. The average time (on my development box) per a replacement was 0.0015 ms for the original method and 0.0362 ms for the new method. Since this it was O(n) and the difference is a fraction of a millisecond, I've decided to run with it for our code base ;)

]]> By: Craig http://www.webdevbros.net/2006/11/20/collection-of-vbscript-string-functions-for-classic-asp/#comment-8792 Craig Thu, 10 Jul 2008 16:27:18 +0000 http://www.webdevbros.net/2006/11/20/collection-of-vbscript-string-functions-for-classic-asp/#comment-8792 Ok, but it's premature optimization without data, so here is some: http://pastebin.com/m34fb4206 It's a csv, so you can download and import into Excel/Calc/whatever. I already identified the most expensive operation, creating the RegExp object. After moving the RegExp out of the method scope and into a global, the new method actually performs better in most cases, and the same in one case. Here's the performance code: http://pastebin.com/m6fb6609d Both of the pastebins will last a month. I'm not sure if I understand the special character wrapping solution. Perhaps an example would help? Ok, but it's premature optimization without data, so here is some:
http://pastebin.com/m34fb4206
It's a csv, so you can download and import into Excel/Calc/whatever.

I already identified the most expensive operation, creating the RegExp object. After moving the RegExp out of the method scope and into a global, the new method actually performs better in most cases, and the same in one case.

Here's the performance code: http://pastebin.com/m6fb6609d

Both of the pastebins will last a month.

I'm not sure if I understand the special character wrapping solution. Perhaps an example would help?

]]> By: Michal http://www.webdevbros.net/2006/11/20/collection-of-vbscript-string-functions-for-classic-asp/#comment-8790 Michal Thu, 10 Jul 2008 15:02:13 +0000 http://www.webdevbros.net/2006/11/20/collection-of-vbscript-string-functions-for-classic-asp/#comment-8790 thx craig ... i am just a bit concerned about performance because this function is used very often in a lot of apps and the change compared to the previous one is dramatic. i am just doing some brainstorming if there would be maybe other solutions as well .. you know some function like this is called a lot. i was thinking of something like: 1. looping through all given placeholders 2. replacing its occurencies and surrounding it with some special char. the special char would prevent the repeated replacement (if regex is done correctly) 3. remove all the special chars. havent tried yet if it would work .. what do you think? if yes, what char could we use? thx craig ... i am just a bit concerned about performance because this function is used very often in a lot of apps and the change compared to the previous one is dramatic.
i am just doing some brainstorming if there would be maybe other solutions as well .. you know some function like this is called a lot.

i was thinking of something like:
1. looping through all given placeholders
2. replacing its occurencies and surrounding it with some special char. the special char would prevent the repeated replacement (if regex is done correctly)
3. remove all the special chars.

havent tried yet if it would work .. what do you think? if yes, what char could we use?

]]> By: Craig http://www.webdevbros.net/2006/11/20/collection-of-vbscript-string-functions-for-classic-asp/#comment-8786 Craig Thu, 10 Jul 2008 13:15:34 +0000 http://www.webdevbros.net/2006/11/20/collection-of-vbscript-string-functions-for-classic-asp/#comment-8786 It's not pretty, but I've come to terms with ASP code never looking pretty ;). I concat Left(...) with the Replace return because the Replace function truncates the string to the left of start. [asp] public function StringFormat(str, args()) if not IsArray(args) then StringFormat = str exit function end if if ubound(args) = -1 then StringFormat = str exit function end if dim regex : set regex = new RegExp regex.Pattern = "{\d+}" regex.Global = true dim match, i, shift shift = 1 for each match in regex.Execute(str) i = CInt(Mid(match.Value, 2, Len(match.Value) - 2)) str = Left(str, match.FirstIndex + shift - 1) & Replace(str, match.Value, args(i), match.FirstIndex + shift, 1, 0) shift = shift + Len(args(i)) - Len(match.Value) next StringFormat = str end function [/asp] It's not pretty, but I've come to terms with ASP code never looking pretty ;). I concat Left(...) with the Replace return because the Replace function truncates the string to the left of start.

PLAIN TEXT
ASP:
  1. public function StringFormat(str, args())
  2.    if not IsArray(args) then
  3.       StringFormat = str
  4.       exit function
  5.    end if
  6.  
  7.    if ubound(args) = -1 then
  8.       StringFormat = str
  9.       exit function
  10.    end if
  11.    
  12.    dim regex : set regex = new RegExp
  13.    regex.Pattern = "{\d+}"
  14.    regex.Global = true
  15.  
  16.    dim match, i, shift
  17.    shift = 1
  18.  
  19.    for each match in regex.Execute(str)
  20.       i = CInt(Mid(match.Value, 2, Len(match.Value) - 2))
  21.       str = Left(str, match.FirstIndex + shift - 1) & Replace(str, match.Value, args(i), match.FirstIndex + shift, 1, 0)
  22.       shift = shift + Len(args(i)) - Len(match.Value)
  23.    next
  24.    StringFormat = str
  25. end function

]]> By: Michal http://www.webdevbros.net/2006/11/20/collection-of-vbscript-string-functions-for-classic-asp/#comment-8785 Michal Wed, 09 Jul 2008 07:58:31 +0000 http://www.webdevbros.net/2006/11/20/collection-of-vbscript-string-functions-for-classic-asp/#comment-8785 craig thats right .. thx for pointing this out... do u have an idea how to solve this nicely? thx for any comments craig thats right .. thx for pointing this out... do u have an idea how to solve this nicely? thx for any comments

]]> By: Craig http://www.webdevbros.net/2006/11/20/collection-of-vbscript-string-functions-for-classic-asp/#comment-8783 Craig Tue, 08 Jul 2008 17:14:59 +0000 http://www.webdevbros.net/2006/11/20/collection-of-vbscript-string-functions-for-classic-asp/#comment-8783 Your Format method does not support the following: Format("{0}{1}", array("{1}", "{0}")). Output should be "{1}{0}" but will be "{0}{0}". Your Format method does not support the following: Format("{0}{1}", array("{1}", "{0}")). Output should be "{1}{0}" but will be "{0}{0}".

]]> By: CC http://www.webdevbros.net/2006/11/20/collection-of-vbscript-string-functions-for-classic-asp/#comment-2383 CC Sat, 02 Feb 2008 12:06:56 +0000 http://www.webdevbros.net/2006/11/20/collection-of-vbscript-string-functions-for-classic-asp/#comment-2383 glitch in stringOperations.asp public function JSEncode(byVal val) val = val & “” tmp = replace(val, chr(39), “\’”) 'it should be here tmp = replace(tmp, chr(92), “\\”) ‘tmp = replace(val, chr(39), “\’”) the order is not correct tmp = replace(tmp, chr(34), “"”) tmp = replace(tmp, chr(13), “”) tmp = replace(tmp, chr(10), ” “) JSEncode = tmp end function better follow the XHTML way to use “hidden” instead of “Hidden” in public function getHiddenInput(name, value) glitch in stringOperations.asp
public function JSEncode(byVal val)
val = val & “”
tmp = replace(val, chr(39), “\’”) 'it should be here
tmp = replace(tmp, chr(92), “\\”)
‘tmp = replace(val, chr(39), “\’”) the order is not correct
tmp = replace(tmp, chr(34), “"”)
tmp = replace(tmp, chr(13), “”)
tmp = replace(tmp, chr(10), ” “)
JSEncode = tmp
end function

better follow the XHTML way to use “hidden” instead of “Hidden” in
public function getHiddenInput(name, value)

]]> By: retro http://www.webdevbros.net/2006/11/20/collection-of-vbscript-string-functions-for-classic-asp/#comment-1137 retro Thu, 01 Nov 2007 14:51:02 +0000 http://www.webdevbros.net/2006/11/20/collection-of-vbscript-string-functions-for-classic-asp/#comment-1137 I do love the classics. I do love the classics.

]]>