replace multiple strings in 100mb file

[DCCD]

Hi, i wrote a script that can replace multiple strings in a xml file works fine but so slow!

I've used StringReplace ,_ReplaceStringInFile, StringRegExpReplace, all the same very slow,.

The number of replacements in the file about 8000

Any help would be greatly appreciated

expandcollapse popup

#include <File.au3>
$path = @ScriptDir & '\xmlfo.xml'
$OXML = FileOpen($path, 256)
$XML = FileRead($OXML)

$term = 'post'
$nofr = 1
Local $aArray = StringRegExp($XML, '(?s)<entry[^>]*>.*?</entry>', 3)
FileClose($OXML)
$XL = $XML
If Not @error Then
    For $i = 0 To UBound($aArray) - 1
        ;get data start
        ;ConsoleWrite ( $aArray[0] &' '&$i& @CRLF)
        $date = StringRegExp($aArray[$i], '(?i)<published>(.*?)</published>', 3)
        If @error Then
            $date = StringRegExp("date err", "(.{33,}?(?:\s)|.+)", 3)
        ElseIf Not @error Then
            ;ConsoleWrite($date[0] & ' ' & $i & @CRLF)
        EndIf
        $kind = StringRegExp($aArray[$i], '(?i)<category>(.*?)</category>', 3)
        If @error Then
            $kind = StringRegExp("kind err", "(.{33,}?(?:\s)|.+)", 3)
            ;ConsoleWrite ( $kind[0] &' '&$i& @CRLF)
        ElseIf Not @error Then
            ;ConsoleWrite ( $kind[0] &' '&$i& @CRLF)
        EndIf
        If $kind[0] = $term And Data(getdate($date[0], 'year'), getdate($date[0], 'month')) = True Then

            _ReplaceStringInFile($path, $aArray[$i], '')

            If Not @error Then
                ;MsgBox(16,'',$XL)
                ConsoleWrite($nofr & ' ' & $i & @CRLF)
                $nofr = $nofr + 1
            EndIf
            ;FileDelete(@ScriptDir & '\XML_output.xml')
            ;FileWrite (@ScriptDir & '\XML_output.xml', StringToBinary ( StringReplace($temp, $aArray[$i], "") , 4) )
        Else
            ConsoleWrite ('err0x0'& @CRLF)
        EndIf
    Next
EndIf

Edited January 30, 2015 by DCCD

[SmOke_N]

Well, you have a huge issue with loading and unloading 100mb's into memory over and over.

Every call to _ReplaceStringInFile opens the file twice.

So... 2 suggestions I can think of.

1.  Ditch _ReplaceStringInFile() and just read the file into memory once, enum each line, keep what you want, remove what you don't (would require a second string to write back to the file, I say a second string because _ArrayDelete ReDims the array every time).

2.  Read the file into chunks and repeat step 1.

Edit:

If this is some type of database script, sqlite would make a lot more sense.

Edited January 30, 2015 by SmOke_N

[kylomas]

DCCD,

See jdelaney's sig for working with XML files directly. 

kylomas

[jguinch]

This : StringRegExp("date err", "(.{33,}?(?:s)|.+)", 3)

and this : StringRegExp("kind err", "(.{33,}?(?:s)|.+)", 3)

has not sense...

Can you post a sample of your XML file, and explain us what exactly you want to replace by what ?

[DCCD]

Thank you all, i really appreciate your help and support  , actually i'm still stuck in this slow loop  I've tried many ways and it isn't working

[SmOke_N]

You're going to have us guess without your code and an example file of what you've tried aren't you ... ?

[DCCD]

You're going to have us guess without your code and an example file of what you've tried aren't you ... ?

 

@SmOke_N, thanks a lot, I'll try something different, thanks again for all your help

[DCCD]

Well, you have a huge issue with loading and unloading 100mb's into memory over and over.

Every call to _ReplaceStringInFile opens the file twice.

So... 2 suggestions I can think of.

1.  Ditch _ReplaceStringInFile() and just read the file into memory once, enum each line, keep what you want, remove what you don't (would require a second string to write back to the file, I say a second string because _ArrayDelete ReDims the array every time).

2.  Read the file into chunks and repeat step 1.

Edit:

If this is some type of database script, sqlite would make a lot more sense.

 

each text string need to be replaced may contain more than 500 characters/numbers.

[SmOke_N]

I'm sorry, I don't see the relevance to your statement/reply.

 

Edit:

This would speed up your script exponentially.

expandcollapse popup

#include <File.au3>
$path = @ScriptDir & '\xmlfo.xml'
$OXML = FileOpen($path, 256)
$XML = FileRead($OXML)
FileClose($OXML)

$term = 'post'
$nofr = 1
Local $aArray = StringRegExp($XML, '(?s)<entry[^>]*>.*?</entry>', 3)

If Not @error Then
    For $i = 0 To UBound($aArray) - 1
        ;get data start
        ;ConsoleWrite ( $aArray[0] &' '&$i& @CRLF)
        $date = StringRegExp($aArray[$i], '(?i)<published>(.*?)</published>', 3)
        If @error Then
            $date = StringRegExp("date err", "(.{33,}?(?:\s)|.+)", 3)
        ElseIf Not @error Then
            ;ConsoleWrite($date[0] & ' ' & $i & @CRLF)
        EndIf
        $kind = StringRegExp($aArray[$i], '(?i)<category>(.*?)</category>', 3)
        If @error Then
            $kind = StringRegExp("kind err", "(.{33,}?(?:\s)|.+)", 3)
            ;ConsoleWrite ( $kind[0] &' '&$i& @CRLF)
        ElseIf Not @error Then
            ;ConsoleWrite ( $kind[0] &' '&$i& @CRLF)
        EndIf
        If $kind[0] = $term And Data(getdate($date[0], 'year'), getdate($date[0], 'month')) = True Then
            
            $XML = StringReplace($XML, $aArray[$i], '')
            ;_ReplaceStringInFile($path, $aArray[$i], '')

            If Not @error Then
                ;MsgBox(16,'',$XL)
                ConsoleWrite($nofr & ' ' & $i & @CRLF)
                $nofr = $nofr + 1
            EndIf
            ;FileDelete(@ScriptDir & '\XML_output.xml')
            ;FileWrite (@ScriptDir & '\XML_output.xml', StringToBinary ( StringReplace($temp, $aArray[$i], "") , 4) )
        Else
            ConsoleWrite ('err0x0'& @CRLF)
        EndIf
    Next
EndIf

Global $ghOpen = FileOpen($path, $FO_UTF8_NOBOM + $FO_OVERWRITE)
FileWrite($ghOpen, $XML)
FileClose($ghOpen)

Here, as suggested before, we are only opening the file, reading the file, and writing to the file 1 time.

Your way, it was opening, reading to memory, writing as many times as the loop was long.

One thing is different, the FileOpen at the bottom of the script, you never told _ReplaceStringInFile how to write the data back to the file, so it was writing it regularly, I added $FO_UTF8_NOBOM strictly because that's how you opened it before in your code example.

So you may want to backup your xml file before using this code (just FYI).

Edited February 1, 2015 by SmOke_N

[DCCD]

I'm sorry, I don't see the relevance to your statement/reply.

 

Edit:

This would speed up your script exponentially.

expandcollapse popup

#include <File.au3>
$path = @ScriptDir & '\xmlfo.xml'
$OXML = FileOpen($path, 256)
$XML = FileRead($OXML)
FileClose($OXML)

$term = 'post'
$nofr = 1
Local $aArray = StringRegExp($XML, '(?s)<entry[^>]*>.*?</entry>', 3)

If Not @error Then
    For $i = 0 To UBound($aArray) - 1
        ;get data start
        ;ConsoleWrite ( $aArray[0] &' '&$i& @CRLF)
        $date = StringRegExp($aArray[$i], '(?i)<published>(.*?)</published>', 3)
        If @error Then
            $date = StringRegExp("date err", "(.{33,}?(?:\s)|.+)", 3)
        ElseIf Not @error Then
            ;ConsoleWrite($date[0] & ' ' & $i & @CRLF)
        EndIf
        $kind = StringRegExp($aArray[$i], '(?i)<category>(.*?)</category>', 3)
        If @error Then
            $kind = StringRegExp("kind err", "(.{33,}?(?:\s)|.+)", 3)
            ;ConsoleWrite ( $kind[0] &' '&$i& @CRLF)
        ElseIf Not @error Then
            ;ConsoleWrite ( $kind[0] &' '&$i& @CRLF)
        EndIf
        If $kind[0] = $term And Data(getdate($date[0], 'year'), getdate($date[0], 'month')) = True Then
            
            $XML = StringReplace($XML, $aArray[$i], '')
            ;_ReplaceStringInFile($path, $aArray[$i], '')

            If Not @error Then
                ;MsgBox(16,'',$XL)
                ConsoleWrite($nofr & ' ' & $i & @CRLF)
                $nofr = $nofr + 1
            EndIf
            ;FileDelete(@ScriptDir & '\XML_output.xml')
            ;FileWrite (@ScriptDir & '\XML_output.xml', StringToBinary ( StringReplace($temp, $aArray[$i], "") , 4) )
        Else
            ConsoleWrite ('err0x0'& @CRLF)
        EndIf
    Next
EndIf

Global $ghOpen = FileOpen($path, $FO_UTF8_NOBOM + $FO_OVERWRITE)
FileWrite($ghOpen, $XML)
FileClose($ghOpen)

Here, as suggested before, we are only opening the file, reading the file, and writing to the file 1 time.

Your way, it was opening, reading to memory, writing as many times as the loop was long.

One thing is different, the FileOpen at the bottom of the script, you never told _ReplaceStringInFile how to write the data back to the file, so it was writing it regularly, I added $FO_UTF8_NOBOM strictly because that's how you opened it before in your code example.

So you may want to backup your xml file before using this code (just FYI).

 

@SmOke_N, Thank you for all your help   and I apologize for the late response