[RESOLVED] Question about the speed. See code.

[bPrice]

Private Sub cmdLoad_Click()
    Dim str As String
    Dim strA() As String
    Dim strB() As String
    Dim i As Long, j As Long

    '------------------------------WHY THIS ONE IS SPEEDY? FILE SIZE 20MB------------------------


    Open App.Path & "\raw.txt" For Input As #1
    str = StrConv(InputB(LOF(1), 1), vbUnicode)
    Close #1

    '------------------------------WHY THIS ONE IS SPEEDY? So many loops and functions-----------------------

    strA = Split(str, vbCrLf)
    j = 0
    ProgressBar.Max = UBound(strA) + 1
    ProgressBar.Min = 0
    ProgressBar.Value = 0
    For i = 0 To UBound(strA)
        DoEvents
        If InStr(1, strA(i), "DELIMITER") Then
            ReDim Preserve strB(j)
            strB(j) = strA(i)
            j = j + 1
        End If
        ProgressBar.Value = ProgressBar.Value + 1
    Next


    '------------Why this is SLOW, looks not much difference brom the previous section ----------------------

    ProgressBar.Max = UBound(strB) + 1
    ProgressBar.Min = 0
    ProgressBar.Value = 0
    For i = 0 To UBound(strB)
        DoEvents
        str = str & strB(i) & vbCrLf
        ProgressBar.Value = ProgressBar.Value + 1
    Next

    '---------------------- Haven't tested the following -----------
    Open App.Path & "\rawR.txt" For Output As #2
    Print #2, str
    Close #2

End Sub

[bPrice]

   '------------Why this is SLOW, looks not much difference brom the previous section ----------------------

    ProgressBar.Max = UBound(strB) + 1
    ProgressBar.Min = 0
    ProgressBar.Value = 0
    For i = 0 To UBound(strB)
        DoEvents
        str = str & strB(i) & vbCrLf
        ProgressBar.Value = ProgressBar.Value + 1
    Next

I tried this and it runs at a heaven's speed ...

    str = Join(strB, vbCrLf)

[DataMiser]

Those pieces of code in the OP are all doing different things, You say two are speedy and one is slow but what does that even mean and what are they doing really?

DoEvents can slow things down as well as introduce potential problems. Refresh is a better choice in such a case.

[bPrice]

Those pieces of code in the OP are all doing different things, You say two are speedy and one is slow but what does that even mean and what are they doing really?

DoEvents can slow things down as well as introduce potential problems. Refresh is a better choice in such a case.

1, I have a text file. Now I use Open and InputB function, to load it into a string variable. This is fast even though the file is 20mb.

2, After this, I split this string into different parts (20000+) and according to a certain standard, I then load the string parts into a string array, and this is fast, too.

3, Last I use For Loop, to save each element in this string array to one string variable. This is where things get very slow. I mean, 100 times slower. However, during the whole process, the data in total didn't increase. It's just the ways I handled the data are different.

  '------------Why this is SLOW, looks not much difference brom the previous section ----------------------

    ProgressBar.Max = UBound(strB) + 1
    ProgressBar.Min = 0
    ProgressBar.Value = 0
    For i = 0 To UBound(strB)
        DoEvents
        str = str & strB(i) & vbCrLf
        ProgressBar.Value = ProgressBar.Value + 1
    Next

Later I tried to use JOIN() and it's also fast, 100 times more faster. I'm looking for explanations.

[Arnoutdv]

This makes the code slow:

str = str & strB(i) & vbCrLf

Every time the string needs to be reallocated to add 2 parts of data.
More efficient would be the calculate the total size of the target string, allocate this size to the string and then use the mid() method to update the string.
Or you could use some external stringbuilder class for this.

[Navion]

Are you really using a raw.txt file with the word "DELIMITER" written in it , say, thousands of times on each line?

If not (I doubt so), the if statement :

If InStr(1, strA(i), "DELIMITER") Then

will never evaluate true, no Redim Preserve on strB(j) will ever take place. That in itself would explain the speed of the first loop.

[MarkT]

This explains what Arnoutdv was saying in further detail. https://support.microsoft.com/kb/170964?wa=wsignin1.0
Bottom line is concatination is very slow. The mid method or use of an array are the ways to go.

[techgnome]

Yeah, it's the string concatenation that's the chokepoint. Strings are immutable, so when you append strings it has to re-allocate new memory and copy everything over to the new space. Using the replace method as suggested by Arnoutdv is a better way to go because it doesn't change the size, making allocation a lot faster. I think new strings still get created, but it's still better that straight up concatenation. My instinct was to use the join method ... so I'd compare the join operation with Arnoutdv's suggestion and see how they compare.

-tg