VBspeed: IsSameText

VBspeed / Language / IsSameText

IsSameText

VB5/6 provides the StrComp function for textual comparison of two strings, and it does a good job in string sorting business. But when you just want to know whether the strings are identical or not, is StrComp the weapon of choice?
IsSameText is defined in analogy to StrComp, however it returns a boolean:

Function IsSameText(String1 As String, _
                    String2 As String, _
                    Compare As VbCompareMethod) As Boolean

Note that IsSameText is not a StrComp emulation. StrComp is just the best thing native VB offers for the IsSameText job. See notes.

Code
StrComp	fRet = (StrComp(sDum1, sDum2, vbTextCompare) = 0)
lstrcmpi	Public Declare Function lstrcmpi Lib "kernel32" Alias "lstrcmpiA" (ByVal lpString1 As String, ByVal lpString2 As String) As Long fRet = (lstrcmpi(sDum1, sDum2) = 0)
IsSameText01	Public Function IsSameText01(String1 As String, _ String2 As String, _ Compare As VbCompareMethod) As Boolean ' by Donald, donald@xbeat.net, 20001012, rev 001 20040813 If LenB(String1) = LenB(String2) Then If LenB(String1) = 0 Then IsSameText01 = True Else ' InStrB does not work with vbTextCompare IsSameText01 = (InStr(1, String1, String2, Compare) <> 0) End If End If End Function
IsSameText02	Public Function IsSameText02(String1 As String, _ String2 As String, _ Compare As VbCompareMethod) As Boolean ' by Donald, donald@xbeat.net, 20001012, rev 001 20040813 If LenB(String1) = LenB(String2) Then ' take the best of each worlds If Compare = vbBinaryCompare Then If LenB(String1) = 0 Then IsSameText02 = True Else IsSameText02 = (InStrB(1, String1, String2, Compare) <> 0) End If Else IsSameText02 = (StrComp(String1, String2, Compare) = 0) End If End If End Function
IsSameText03	IsSameText03 is wrapped in a class. Cinemascope here, preview here: see <a href="cod_IsSameText03.htm">cod_IsSameText03.htm</a>
Calls
	Compare = vbTextCompare
1	"xxx..." vs. "XXX...", both strings 1000 chars, all chars equal textwise
2	"xxx..." vs. "YXX...", both strings 1000 chars, first char differs
3	"xxx...x" vs. "XXX...Y", both strings 1000 chars, last char differs
4	"xxx..." vs. "XXX...", 1000 vs. 999 chars, all chars equal textwise
Charts

VB5 Charts

Code	Author	Doping	Notes
StrComp	Donald
lstrcmpi	Jason	API
IsSameText01	Donald
IsSameText02	Donald
IsSameText03	Chris	API

Call 1
3	11.09	371µs
2	5.74	192µs
5	16.83	563µs
4	11.10	371µs
1	1.00	33µs

Call 2
2	408.81	123.114µs
4	484.64	145.953µs
5	1,840.04	554.138µs
3	410.29	123.562µs
1	1.00	0.301µs

Call 3
3	11.16	373µs
2	5.76	193µs
5	16.86	563µs
4	11.16	373µs
1	1.00	33µs

Call 4
5	2,884.02	370.889µs
4	1,487.03	191.234µs
2	1.02	0.131µs
1	1.00	0.129µs
3	1.40	0.180µs

VB6 Charts

Code	Author	Doping	Notes
StrComp	Donald
lstrcmpi	Jason	API
IsSameText01	Donald
IsSameText02	Donald
IsSameText03	Chris	API

Call 1
3	15.94	533µs
2	5.73	192µs
5	17.28	578µs
4	15.96	534µs
1	1.00	33µs

Call 2
3	970.67	283.917µs
2	496.93	145.350µs
5	1,893.93	553.967µs
4	972.28	284.387µs
1	1.00	0.292µs

Call 3
3	15.96	533µs
2	5.75	192µs
5	17.31	578µs
4	15.99	534µs
1	1.00	33µs

Call 4
5	5,228.01	571.065µs
4	1,755.31	191.736µs
2	1.12	0.123µs
1	1.00	0.109µs
3	1.49	0.163µs

Conclusions
Comparing IsSameText01 with IsSameText02 the former loses everywhere so we can just forget about it. Learn: InStr is not as good with textual comparison as it is with binary comparison (cf. Same String?). IsSameText02 now looks a clear winner: although it involves a function call while the rivaling StrComp is in-line there's hardly any difference with same-length strings (Call 1 to 3), and with different-length strings IsSameText02 beats the hell out of StrComp (Call 4). I mean, 3000 to 5000 times faster sounds like convincing... Note, how StrComp got considerably slower in VB6. 11-Sep-2001: the lstrcmpi API is clearly best with same-length strings. Good to know. 06-Dec-2001: IsSameText03 happened ... looks like a real alternative. The bottom line is: Use IsSameText03 to find out whether two strings are textually identical. The bottom line was: Use IsSameText02 to find out whether two strings are textually identical. By the way: IsSameText02 is also a good choice for binary comparison.
NOTES
Note that IsSameText is not about emulating StrComp! StrComp is an animal of sorting, and thus has to deal with the irregularities of human language orthography and lexical sorting standards. For example (InStr can do this trick, too): StrComp("Masse", "Maße", vbTextCompare) --> 0 (equal) InStr(1, "Masse", "Maße", vbTextCompare) --> 1 However, check out this one: InStr(1, "as", "aß", vbTextCompare) --> 0 (not found: 'aß'='ass' is not in 'as') InStr(1, "aß", "as", vbTextCompare) --> 1 (found as pos 1: 'as' is in 'aß'='ass') So, in textual comparison mode, "ß" is first transliterated to "ss", and then the search begins. The same holds for all characters with an alternative digraphemic transliteration: "ß" = "SS","Ss","sS","ss" "Œ","œ" = "OE","Oe","oE","oe" "Æ","æ" = "AE","Ae","aE","ae" "Þ","þ" = "TH","Th","tH","th"
Making notes on StrComp, here's a last one: VB5's StrComp sorts the number characters "1", "2", "3" (Ascii 49,50,51) before their superscript counterparts, whereas VB6 does not make any difference: VB5 StrComp: 1(49) < ¹(185), 2(50) < ²(178), 3(51) < ³(179) VB6 StrComp: 1(49) = ¹(185), 2(50) = ²(178), 3(51) = ³(179)
Got comments?

top