VB Excel Deleting Rows On A Specified Value Being In a Cell

Here are two common patterns for deleting entire rows based on a condition. The main idea is that you cannot delete from a collection while you iterate it. This means that Delete should not appear in a For Each This is fairly standard across most programming languages and some even throw an error to prevent it.

Option 1, use an integer to track the rows and have it work from the end to the beginning. You need to go backwards because it is the easy way to avoid skipping rows. It is possible to go forwards, you just need to account for not incrementing when you delete.

Sub DeleteRowsWithIntegerLoop()

    Dim rng_delete As Range
    Set rng_delete = Range("A1:A1000")

    Dim int_start As Integer
    int_start = rng_delete.Rows.Count

    Dim i As Integer
    For i = int_start To 1 Step -1
        If InStr(rng_delete.Cells(i), "VALUE") > 0 Then
            rng_delete.Cells(i).EntireRow.Delete
        End If
    Next i

End Sub

Option 2, use the Union-Delete pattern to build a range of cells and then delete them all in one step at the end.

Sub DeleteRowsWithUnionDelete()


    Dim rng_cell As Range
    Dim rng_delete As Range

    For Each rng_cell In Range("A1:A1000")
        If InStr(rng_cell, "VALUE") > 0 Then
            If rng_delete Is Nothing Then
                Set rng_delete = rng_cell
            Else
                Set rng_delete = Union(rng_delete, rng_cell)
            End If
        End If
    Next

    rng_delete.EntireRow.Delete

End Sub

Notes on the code

For Option 2, there is an extra conditional there to create rng_delete when it starts at the first item. Union does not work with a Nothing reference, so we first check that and if so, Set to the first item. All others come through and get Set by the Union line.

Preference

When choosing between the two, I always prefer Option 2 because I much prefer to work with Ranges in Excel instead of iterating through Cells with a counter. There are limitations to this. The second option also works for discontinuous Ranges and all other variety of weird Ranges (e.g. after a call to SpecialCells) which can make it valuable when you are not sure what data you will be dealing with.

Speed

I am not sure about a speed comparison. Both can be slow if ScreenUpdating and calculations are enabled. The first option makes N-1 calls to Delete whereas the second option does a single one. Delete is an expensive operation. Option 2 does however make N-1 calls to Union and Set. I assume it's faster than the first one based on that (and it seems to be here), but I did not profile it.

Final note: InStr returns an integer indicating where the value was found. I always like to make the boolean covnersion explicit here and compare to >0.

Dont use loops, use a method such as AutoFilter

Sub Recut()
Dim ws As Worksheet
Dim rng1 As Range

Set ws = ActiveSheet
Set rng1 = ws.[A1:A100]

ws.AutoFilterMode = False

Application.ScreenUpdating = False
With rng1
      .AutoFilter Field:=1, Criteria1:="*VALUE*"
      `assuming A1 is a header, and checking for at least 1 match on the filtered string
       If .SpecialCells(xlCellTypeVisible).Cells.Count > 1 Then .Offset(1, 0).Resize(rng1.Rows.Count - 1, 1).EntireRow.Delete
End With
ws.AutoFilterMode = False
Application.ScreenUpdating = True

End Sub