_StringSplit without ubound in [0], faster?

randallc · January 23, 2008

Hi,

There is a thread in "request forum", showing that an option for "StringSplit" without the extra count element at the beginning is "not going to happen"

I have thought about how to get it faster, given that _ArrayDelete($ar,0) is so slow;

Here is my second suggestion!

OMG! wot wos i thinking...

Func _StringSplit2($sStr, $sSepF, $iFlag = 0)
    Local $sTmp = $sStr, $arFileF[1] = [$sStr], $iPos, $iPos1, $sLenSep = StringLen($sSepF), $iCaseSens = 1
    If $sSepF = "" Then ;handle empty splitter separator
        $iPos = 2
        $sTmp = StringLeft($sStr, 1)
        ; if Sep has more than 1 character, we must first check for ;split the Separator into individual characters
    ElseIf Not $iFlag And $sLenSep > 1 Then ;maybe we should use RegExp
        ; check the first character as separator, to find likely $iPos candidate
        Local $aSep = StringSplit($sSepF, "", 1) ;split the Separator into individual characters
        $iPos = StringInStr($sTmp, $aSep[1], $iCaseSens)
        $sTmp = StringLeft($sTmp, $iPos - 1) ;keep as short as possible
        For $c = 2 To UBound($aSep) - 1; loop through the rest of the separators in case any are sooner
            $iPos1 = StringInStr($sTmp, $aSep[$c ], $iCaseSens)
            If $iPos1 And $iPos1 < $iPos Then ;must be found, and < $iPos
                $iPos = $iPos1
                $sTmp = StringLeft($sTmp, $iPos - 1) ;keep as short as possible
            EndIf
        Next
    Else ;if $iFlag=1 or only 1 character, above not needed
        If $sLenSep > 1 Then $iCaseSens = 0 ;stringsplit is NOT case-sensitive if more than 1 char in separator, used as 1
        $iPos = StringInStr($sStr, $sSepF, $iCaseSens)
        $sTmp = StringLeft($sStr, $iPos - 1)
    EndIf
    If Not $iPos Then Return SetError(@error, 0, $arFileF)
    If Not $iFlag And $sLenSep Then $sLenSep = 1 ;change length for single char separators of "0" $iFlag
    $sStr = StringMid($sStr, $iPos + $sLenSep)
    $arFileF = StringSplit($sStr, $sSepF, $iFlag)
    $arFileF[0] = $sTmp
    Return SetError(0, 0, $arFileF);$arFileF
EndFunc   ;==>_StringSplit2

Modified due to:

1. Bug for 2 elements.

2. Case anomaly in "StringSplit"

3. Case of more than 1 character in Separator not handled except by group

4. Empty separator not handled accurately.

Any feedback welcome asap, as the func in last beta was buggy...

Thanks, Randall

Edited February 14, 2008 by randallc

Thatsgreat2345 · January 23, 2008

Hi,

There is a thread in "request forum", showing that an option for "StringSplit" without the extra count element at the beginning is "not going to happen"

I have thought about how to get it faster, given that _ArrayDelete($ar,0) is so slow;

Here is my first suggestion, using RegExp (about 3x as fast as inbuilt), but it needs fixing;

$sStr = $sSepF & $sStr & $sSepF
$sPattern = "(?=[^|" & $sSepF & "])([^" & $sSepF & "]*)(?=" & $sSepF & ")"
$arFileF = StringRegExp($sStr, $sPattern, 3)oÝ÷ ÛSh¶·®'¯x©äzW¦z{ljÚkzÇ§·c^yÛn)b·+h¥«,é®)íx,¦Xr«iË^®À'Ê¥¦è¶«²è zËb¢{¢·Ú²×«jk¡§!Â¥u·¦Þr&yÝ÷ë-E©ÝjYZºÚ"µÍÈÝ[ÔÜ]]LÂÚ[ÛYIØ^K]LÉÝÂßÐ]]Ò]ÕÜÕÙP[ÚO^BÌÍÜÔÙH  ][ÝË  ][ÝÂßØØ[   ÌÍÜÑ[HH[SÜ[X[ÙÊ  ][ÝÉ][ÝËØÜ    ][ÝÊ

I][ÝÊK    ÌÍÜÑ[XYHÝ[ÔXÙJ[TXY
    ÌÍÜÑ[JKÔ   ][ÝË  ][ÝÊBßØØ[  ÌÍØ[HHÝ[ÔÜ]
    ÌÍÜÑ[XY ][ÝË  ][ÝÊK ÌÍØÈHÐ^Q[]J    ÌÍØ[K
BßØØ[    ÌÍØ[HHÝ[ÔÜ]
    ÌÍÜÑ[XYÔJK ÌÍØÈHÐ^Q[]J    ÌÍØ[K
BØØ[  ÌÍÜÑ[XYH    ][ÝØKËKËKKËKËKKËKËKKËKËXKËKËKKËKËI][ÝÂÌÍØ[HHÔÝ[ÔÜ]J  ÌÍÜÑ[XY ÌÍÜÔÙ
BßÐ^QÜ^J ÌÍØ[K    ][ÝÉÌÍØ[I][ÝÊBÏOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOBÌÍØ[LHHÔÝ[ÔÜ]
    ÌÍÜÑ[XY ÌÍÜÔÙ
BßÐ^QÜ^J ÌÍØ[LK   ][ÝÉÌÍØ[LI][ÝÊBÏOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOB[ÈÔÝ[ÔÜ]
    ÌÍÜÔÝ  ÌÍÜÔÙ  ÌÍÚS[HHL
NÈ ][ÝÉ[ÉÌÍÜÔÙ [É][ÝÂSØØ[ ÌÍÜÕ[YH[Y[]

BQÜ    ÌÍÚHHÈ  ÌÍÚS[BBSØØ[    ÌÍØ[QHÝ[ÔÜ]
    ÌÍÜÔÝ  ÌÍÜÔÙJK    ÌÍØÈHÐ^Q[]J    ÌÍØ[Q
BS^PÛÛÛÛUÜ]J   ][ÝÝ[YLI][ÝÈ    [ÈÝ[
[YY ÌÍÜÕ[YKH    [ÈBPÛÛÛÛUÜ]J  ][ÝÕPÝ[
    ÌÍØ[QOI][ÝÈ    [ÈPÝ[
    ÌÍØ[QH   [ÈBPÛÛÛÛUÜ]J  ][ÝÉÌÍØ[QÌOI][ÝÈ    [È ÌÍØ[QÌH [ÈBPÛÛÛÛUÜ]J  ][ÝÉÌÍØ[QÕPÝ[
    ÌÍØ[QKLWOI][ÝÈ [È ÌÍØ[QÕPÝ[
    ÌÍØ[QHHWH    [ÈBT]  ÌÍØ[Q[[ÈÏOIÝ×ÔÝ[ÔÜ][ÈÔÝ[ÔÜ]J    ÌÍÜÔÝ  ÌÍÜÔÙ  ÌÍÚS[HHL
NÈ ][ÝÉ[ÉÌÍÜÔÙ [É][ÝÂSØØ[ ÌÍÜÕ[YH[Y[]

BSØØ[ ÌÍÜÔ]H  ][ÝÊÏV× ][ÝÈ  [È ÌÍÜÔÙ  [È ][Ý×JJ×][ÝÈ    [È ÌÍÜÔÙ  [È ][Ý×JJÏI][ÝÈ   [È ÌÍÜÔÙ  [È ][ÝÊI][ÝÂPÛÛÛÛUÜ]J ][ÝÉÌÍÜÔ]I][ÝÈ  [È ÌÍÜÔ]   [ÈBIÌÍÜÔÝH    ÌÍÜÔÙ  [È ÌÍÜÔÝ  [È ÌÍÜÔÙQÜ   ÌÍÚHHÈ  ÌÍÚS[BBIÌÍØ[QHÝ[ÔYÑ^
    ÌÍÜÔÝ  ÌÍÜÔ]ÊBS^PÛÛÛÛUÜ]J    ][ÝÝ[YLI][ÝÈ    [ÈÝ[
[YY ÌÍÜÕ[YKH    [ÈBPÛÛÛÛUÜ]J  ][ÝÕPÝ[
    ÌÍØ[QOI][ÝÈ    [ÈPÝ[
    ÌÍØ[QH   [ÈBPÛÛÛÛUÜ]J  ][ÝÉÌÍØ[QÌOI][ÝÈ    [È ÌÍØ[QÌH [ÈBPÛÛÛÛUÜ]J  ][ÝÉÌÍØ[QÕPÝ[
    ÌÍØ[QKLWOI][ÝÈ [È ÌÍØ[QÕPÝ[
    ÌÍØ[QHHWH    [ÈBT]  ÌÍØ[Q[[ÈÏOIÝ×ÔÝ[ÔÜ]

Hmmm looks good. I may give it a try I use stringsplit alot. ( wow quote and edit just mess up everything wtf)

BrettF · January 23, 2008

Hmmm... using this string:

Local $sFileread = "StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|" _ 
& "StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|" _ 
& "StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|" _ 
& "StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|" _ 
& "StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|" _ 
& "StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|" _ 
& "StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|" _ 
& "StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|" _ 
& "StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|" _ 
& "StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|" _ 
& "StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|" _ 
& "StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|" _ 
& "StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|" _ 
& "StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|" _ 
& "StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|" _ 
& "StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|" _ 
& "StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|" _ 
& "StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|" _ 
& "StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|" _ 
& "StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|" _ 
& "StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|" _ 
& "StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|" _ 
& "StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|" _ 
& "StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|" _ 
& "StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|" _ 
& "StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah"

The results are:

$sPattern=(?=[^|,])([^,]*)(?=,)

time2=30.16

UBound($arFileF)=1

$arFileF[0]=StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBl

hBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah

ingBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah

time2=12.73

UBound($arFileF)=1

$arFileF[0]=StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBl

hBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah

ingBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah|StringBlahBlahBlah

Is that good??

randallc · January 23, 2008

Hmmm... using this string:

Hi,

Not sure how to answer that!

I asked for either

1. Help to fix the Regexp

or

2. An alternative suggestion to speed up StringSplit;

Is this under (1.) because it shows (what I already knew) that the regexp needs fixing still (especially for non-alphabetical split characters etc...)?

Best, randall

PS - I guess you could say (1.) also because it shows that splits with very large elements may be slower with RegExp; but not for small elements, or "real life" tests with splitting lines of files, for instance, when I have tested...

Edited January 23, 2008 by randallc

randallc · January 23, 2008

Hi,

OMG! wot wos i thinking...

Func _StringSplit($sStr, $sSepF, $iFlag=0)
    Local $iPos = StringInStr($sStr, $sSepF), $sTmp = StringLeft($sStr, $iPos - 1)
    $sStr = StringMid($sStr, $iPos + StringLen($sSepF))
    Local $arFileF = StringSplit($sStr, $sSepF, $iFlag)
    $arFileF[0] = $sTmp
    Return $arFileF
EndFunc   ;==>_StringSplit

Edited January 23, 2008 by randallc

BrettF · January 23, 2008

Hi,

OMG! wot wos i thinking...

Func _StringSplit($sStr, $sSepF, $iFlag=0)
    Local $iPos = StringInStr($sStr, $sSepF), $sTmp = StringLeft($sStr, $iPos - 1)
    $sStr = StringMid($sStr, $iPos + StringLen($sSepF))
    Local $arFileF = StringSplit($sStr, $sSepF, $iFlag)
    $arFileF[0] = $sTmp
    Return $arFileF
EndFunc   ;==>_StringSplit

Who knows...

Looks good!

Thatsgreat2345 · January 23, 2008

I kinda prefer it starting at 1 , since I'm so used to just doing for $i = 1 to Ubound($array) - 1 .

randallc · February 14, 2008

Hi,

Testing if possible? - post #1 has new protype, about 4x faster on decent size stringsplit than using ArrayDelete.

Modified due to:

1. Bug for 2 elements.

2. Case anomaly in "StringSplit"

3. Case of more than 1 character in Separator not handled except by group

4. Empty separator not handled accurately.

Any feedback welcome asap, as the func in last beta was buggy...

Thanks, Randall

[PS inbuilt StringSplit found, on testing, to be Case-Sensitive, but not on multiple character separator (when flag=1, "treat as group, not individual chars"); is this deliberate, do you think!?]

Edited February 14, 2008 by randallc

Sign In

_StringSplit without ubound in [0], faster?

Recommended Posts

randallc

Link to comment

Share on other sites

Thatsgreat2345

Link to comment

Share on other sites

BrettF

Link to comment

Share on other sites

randallc

Link to comment

Share on other sites

randallc

Link to comment

Share on other sites

BrettF

Link to comment

Share on other sites

Thatsgreat2345

Link to comment

Share on other sites

randallc

Link to comment

Share on other sites

Create an account or sign in to comment

Create an account

Sign in

Recently Browsing 0 members

Browse

AutoIt Resources

Release

Beta