Strings & math! help me out guys

Something like this?

VB Code:

Sub SubStrings(strFirst As String, strSecond As String)
 
'counter(i) will contain the no. of times a substring
'length i of strSecond appears in strFirst
 
Dim tmparr() As String
Dim tmpStr As String
Dim Counter() As Long
Dim i As Long, j As Long
 
ReDim Counter(Len(strSecond) )
 
For i = 1 To Len(strSecond)
    For j = 1 To Len(strSecond) - i + 1
        tmpStr = Mid$(strSecond, j, i)
        tmparr() = Split(strFirst, tmpStr, , vbBinaryCompare)
        Counter(i) = UBound(tmparr()) + Counter(i)
    Next j
Next i
 
End Sub

Ok, I'll take a shot.

I have a function DelimitedString_NumberofStrings that will
retrieve the number of pieces of strings that are delimited by a
constant string:

xxx = "Thefoxwalksonthefloor"

DS_NumberOfStrings(xxx,"the") will return 3
and there are 3 fields:"","foxwalkson",and "floor"

Assuming that you have this function

Str1 ="Thefoxwalksonthefloor"
Str2="Samanthafoxfloorsthewalker"
i = 1
hits = 0
while i < (len(STR2)-2)
checkit=mid$(STR2,i,3)
hits = hits+DS_NumberOfStrings(STR1,checkit)
i = i + 1
wend

should do the trick.

Would you like the function list?

Good Luck

hi guys,

Thanks for the good work, but....:eek:

i think i didnt explain it correctly Let me try again. Its a mathematical question;

Str1 ="Thefoxwalksonthefloor"
Str2="Samanthafoxfloorsthewalker"

I want to check for how many 'percent' Str1 equals Str2
(well its not actually 'percents' but i want: The higher the number to bigger the equality.

I use a method where i delimit the strings in 2 character strings and compare them (wita 2 arrays etc). every time i get a match i add 1 to a counter. (that works fine)

Now, To get a better result i ALSO want to compare the strings with 3-caracter strings (the same way as above, but then with 3 -char. strings :) )

The chance that i hit a match with the 3-char. test is far smaller then with the 2-char test. so a match with the 3-char. match "weighs" heavier then th eother one, so i have to add more then 1 to Counter when i have a 3-char match.

My question is: how much more "heavy" is a 3-char match?


Since my native tongue is not english i dont know all the official mathematical language, but i hope u guys understand :p

Thanks !!

Remvs
Holland

Derfarm,

I'd really appreciate the func.list so i can compare a little.


i'm a novice at this and probably have messed up my declarations etc. Since i have to compare 2 x 28000 strings (long ones) im also looking for ways to speed the basterd up a little :D


thnx in advance
remvs

i dont have the code here now, but i can post it tommorow from work if anyone wants to see it

remvs

VB Code:

Sub SubStrings(strFirst As String, strSecond As String)
 
'percentcorrect(i) will contain the % of substring of
'length i of strSecond is succesfully matched with a substring of
'strFirst
 
Dim tmparr() As String
Dim tmpStr As String
Dim Counter As Long
Dim i As Long, j As Long
Dim percentcorrect() As Long
 
ReDim percentcorrect(Len(strSecond))
 
For i = 1 To Len(strSecond)
    For j = 1 To Len(strSecond) - i + 1
        tmpStr = Mid$(strSecond, j, i)
        tmparr() = Split(strFirst, tmpStr, , vbBinaryCompare)
        Counter = UBound(tmparr()) + Counter(i)
    Next j
    percentcorrect(i) = 100 * Counter /(Len(strSecond) - i)
Next i
 
End Sub

Also, assuming you only use letters in your string, a 3 chr string is 26 / no. of comparisons more likely to occur.

ie, if you have 2 * 4chr length strings,

prob. of finding 1 match for a 2chr substring is: 3 * 3 / 26^2
prob. of finding 1 match for a 3chr substring is: 2 * 2 / 26^3

General formula for the probability of 1 substring (length x), of String 1 (length y) occuring in String 2 (length z) is:

(y + 1 - x) * (z + 1 - x) / 26^x

Chrisf,

The formula at the bottom was what i needed! with that i can calculate the probability.

Thnx (all of u guys)

Remvs

You can expand it to find the prob. of n substrings ( length x )
of string 1 occuring in string 2. It will be a Geometric series,
First term n= (y + 1 - x) * (z + 1 - x) / 26^x
Ration = (y + 1 - x)

Giving: ((y + 1 - x) * (z + 1 - x) / 26^x )* (1 - (y + 1 - x)^ n) / (y - x)

if I've got the correct ratio.