UnNest indefinite number of nested objects in vba - vba

I would like to take any number of objects via a ParamArray and then add them, or variables nested within them to a collection. The tricky part is that if that nested object is a container of some sort (collection, scripting dictionary or even a custom class with a count method) also has variables nested within it, I want it to return those in the collection, NOT the container.
It would go something like this, let's start by creating a use case:
Sub MakeItems()
Dim ReturnedColl as Collection
Dim aString as String
Dim TopColl as New Collection, NestedColl as New Collection, SubNestedDic as New Dictionary
Dim aRangeofManyCells as Range, aRangeofOneCell as Range
Dim anObject as newObject, NestedObject as New Object, SubNestedObject as New Object
aString = "Just a string"
Set aRangeofManyCells = Range("A1:C3")
Set aRangeofOneCell = Range("A4")
SubNestedDic.Add SubNestedObject
SubNestedDic.Add aRangeofOneCell
NestedColl.Add SubNestedDic
NestedColl.Add NestedObject
NestedColl.Add SubNestedDic
NestedColl.Add aRangeofManyCells
TopColl.Add aString
TopColl.AddNestedColl
Set ReturnedColl = UnNest(TopColl, TopColl, anObject, Range("Sheet1:Sheet3!Q1"))
For each Item in ReturnedColl
'do something
Next Item
End Sub
Here comes the part I can't figure out.
I would want to do a loop like this making the Item the new Items, and then look into each Item within item (if it has any), but without losing track of the original Items, because I'll have to go to the next Item.
Function UnNest(ParamArray Items() as Variant) as Collection
For Each Item in Items
If Item 'is a container of some sort' Then
'some kind of loop through all nests, subnests, subsubnests,...
Else
UnNest.Add Item
Endif
Next Item
End Function
So the end result should be a collection that holds:
"Just a String" from aString
9 range objects corresponding to the cells Range("A1:C3") from aRangeofManyCells
1 range object corresponding to Range("A4"), from aRangeofOneCell
The objects anObject, NestedObject, and SubNestedObject
All of the above 2x, because I put TopColl as an argument to the Function 2x
And also,
an additional anObject, because I added that as an argument to the function
3 Range objects, corresponding to Sheet1Q1, Sheet2Q2, Sheet3Q3
I know that's a tall order, but there has got to be some way to do that loop.
Thanks for any help!

This routine would appear to solve one of your use cases. Certainly it worked for me although I was not passing anything other than regular variables and arrays.
One problem I could not overcome was that I could not determine the type of an Object. Unless you can solve that problem, I do not see how to achieve your entire objective.
Sub DeNestParamArray(RetnValue() As Variant, ParamArray Nested() As Variant)
' Coded Nov 2010
' Each time a ParamArray is passed to a sub-routine, it is nested in a one
' element Variant array. This routine finds the bottom level of the nesting and
' sets RetnValue to the values in the original parameter array so that other routine
' need not be concerned with this complication.
Dim NestedCrnt As Variant
Dim Inx As Integer
NestedCrnt = Nested
' Find bottom level of nesting
Do While True
If VarType(NestedCrnt) < vbArray Then
' Have found a non-array element so must have reached the bottom level
Debug.Assert False ' Should have exited loop at previous level
Exit Do
End If
If NumDim(NestedCrnt) = 1 Then
If LBound(NestedCrnt) = UBound(NestedCrnt) Then
' This is a one element array
If VarType(NestedCrnt(LBound(NestedCrnt))) < vbArray Then
' But it does not contain an array so the user only specified
' one value; a literal or a non-array variable
' This is a valid exit from this loop
Exit Do
End If
NestedCrnt = NestedCrnt(LBound(NestedCrnt))
Else
' This is a one-dimensional, non-nested array
' This is the usual exit from this loop
Exit Do
End If
Else
Debug.Assert False ' This is an array but not a one-dimensional array
Exit Do
End If
Loop
' Have found bottom level array. Save contents in Return array.
ReDim RetnValue(LBound(NestedCrnt) To UBound(NestedCrnt))
For Inx = LBound(NestedCrnt) To UBound(NestedCrnt)
If VarType(NestedCrnt(Inx)) = vbObject Then
Set RetnValue(Inx) = NestedCrnt(Inx)
Else
RetnValue(Inx) = NestedCrnt(Inx)
End If
Next
End Sub
Public Function NumDim(ParamArray TestArray() As Variant) As Integer
' Returns the number of dimensions of TestArray.
' If there is an official way of determining the number of dimensions, I cannot find it.
' This routine tests for dimension 1, 2, 3 and so on until it get a failure.
' By trapping that failure it can determine the last test that did not fail.
' Coded June 2010. Documentation added July 2010.
' * TestArray() is a ParamArray because it allows the passing of arrays of any type.
' * The array to be tested in not TestArray but TestArray(LBound(TestArray)).
' * The routine does not validate that TestArray(LBound(TestArray)) is an array. If
' it is not an array, the routine return 0.
' * The routine does not check for more than one parameter. If the call was
' NumDim(MyArray1, MyArray2), it would ignore MyArray2.
Dim TestDim As Integer
Dim TestResult As Integer
On Error GoTo Finish
TestDim = 1
Do While True
TestResult = LBound(TestArray(LBound(TestArray)), TestDim)
TestDim = TestDim + 1
Loop
Finish:
NumDim = TestDim - 1
End Function

Related

EXCEL VBA how to use functions and split to extract integer from string

I'm working on a piece of code to extract the nominal size of a pipeline from it's tagname. For example: L-P-50-00XX-0000-000. The 50 would be it's nominal size (2") which I would like to extract. I know I could do it like this:
TagnameArray() = Split("L-P-50-00XX-0000-000", "-")
DNSize = TagnameArray(2)
But I would like it to be a function because it's a small part of my whole macro and I don't need it for all the plants I'm working on just this one. My current code is:
Sub WBDA_XXX()
Dim a As Range, b As Range
Dim TagnameArray() As String
Dim DNMaat As String
Dim DN As String
Set a = Selection
For Each b In a.Rows
IntRow = b.Row
TagnameArray() = Split(Cells(IntRow, 2).Value, "-")
DN = DNMaat(IntRow, TagnameArray())
Cells(IntRow, 3).Value = DN
Next b
End Sub
Function DNMaat(IntRow As Integer, TagnameArray() As String) As Integer
For i = LBound(TagnameArray()) To UBound(TagnameArray())
If IsNumeric(TagnameArray(i)) = True Then
DNMaat = TagnameArray(i)
Exit For
End If
Next i
End Function
However this code gives me a matrix expected error which I don't know how to resolve. I would also like to use the nominal size in further calculations so it will have to be converted to an integer after extracting it from the tagname. Does anyone see where I made a mistake in my code?
This is easy enough to do with a split, and a little help from the 'Like' evaluation.
A bit of background on 'Like' - It will return TRUE or FALSE based on whether an input variable matches a given pattern. In the pattern [A-Z] means it can be any uppercase letter between A and Z, and # means any number.
The code:
' Function declared to return variant strictly for returning a Null string or a Long
Public Function PipeSize(ByVal TagName As String) As Variant
' If TagName doesn't meet the tag formatting requirements, return a null string
If Not TagName Like "[A-Z]-[A-Z]-##-##[A-Z]-####-###" Then
PipeSize = vbNullString
Exit Function
End If
' This will hold our split pipecodes
Dim PipeCodes As Variant
PipeCodes = Split(TagName, "-")
' Return the code in position 2 (Split returns a 0 based array by default)
PipeSize = PipeCodes(2)
End Function
You will want to consider changing the return type of the function depending on your needs. It will return a null string if the input tag doesnt match the pattern, otherwise it returns a long (number). You can change it to return a string if needed, or you can write a second function to interpret the number to it's length.
Here's a refactored version of your code that finds just the first numeric tag. I cleaned up your code a bit, and I think I found the bug as well. You were declaring DNMAAT as a String but also calling it as a Function. This was likely causing your Array expected error.
Here's the code:
' Don't use underscores '_' in names. These hold special value in VBA.
Sub WBDAXXX()
Dim a As Range, b As Range
Dim IntRow As Long
Set a = Selection
For Each b In a.Rows
IntRow = b.Row
' No need to a middleman here. I directly pass the split values
' since the middleman was only used for the function. Same goes for cutting DN.
' Also, be sure to qualify these 'Cells' ranges. Relying on implicit
' Activesheet is dangerous and unpredictable.
Cells(IntRow, 3).value = DNMaat(Split(Cells(IntRow, 2).value, "-"))
Next b
End Sub
' By telling the function to expect a normal variant, we can input any
' value we like. This can be dangerous if you dont anticipate the errors
' caused by Variants. Thus, I check for Arrayness on the first line and
' exit the function if an input value will cause an issue.
Function DNMaat(TagnameArray As Variant) As Long
If Not IsArray(TagnameArray) Then Exit Function
Dim i As Long
For i = LBound(TagnameArray) To UBound(TagnameArray)
If IsNumeric(TagnameArray(i)) = True Then
DNMaat = TagnameArray(i)
Exit Function
End If
Next i
End Function
The error matrix expected is thrown by the compiler because you have defined DNMaat twice: Once as string variable and once as a function. Remove the definition as variable.
Another thing: Your function will return an integer, but you assigning it to a string (and this string is used just to write the result into a cell). Get rid of the variable DN and assign it directly:
Cells(IntRow, 3).Value = DNMaat(IntRow, TagnameArray())
Plus the global advice to use option explicit to enforce definition of all used variables and to define a variable holding a row/column number always as long and not as integer

Syntax options creating errors in VBA Macro for Excel

I'm having some trouble with syntax options while writing a VBA Macro for Excel. In VBA you can call a method on an object in two different ways:
foo.bar(arg1, arg2)
or
foo.bar arg1, arg2
I absolutely detest the second sort of syntax because I find it lacks any sort of clarity, so I normally adhere to the first option. However, I've come across a situation where using the first option creates an error, while the second executes fine. (This may perhaps be an indicator of other problems in my code.) Here is the culprit code:
Function GetFundList() As Collection
Dim newFund As FundValues
Range("A5").Select
Set GetFundList = New Collection
While Len(Selection.Value)
Set newFund = New FundValues
' I set the fields of newFund and move Selection
The problem is in this next line:
GetFundList.Add newFund
Wend
End Function
FundValues is a class I created that is essentially just a struct; it has three properties which get set during the loop.
Basically, when I call GetFundList.Add(newFund) I get the following error:
Run-time error '438':
Object doesn't support this property or method
But calling GetFundList.Add newFund is perfectly fine.
Does anyone understand the intricacies of VBA well enough to explain why this is happening?
EDIT: Thanks much for the explanations!
Adding items to a collection is not defined as a function returning a value, but as a sub routine:
Public Sub Add( _
ByVal Item As Object, _
Optional ByVal Key As String, _
Optional ByVal { Before | After } As Object = Nothing _
)
When calling another sub routine by name and sending arguments (without adding the "Call" statement), you are not required to add parentheses.
You need to add parentheses when you call a function that returns a value to a variable.
Example:
Sub Test_1()
Dim iCnt As Integer
Dim iCnt_B As Integer
Dim iResult As Integer
iCnt = 2
iCnt_B = 3
fTest_1 iCnt, iResult, iCnt_B
End Sub
Public Function fTest_1(iCnt, iResult, iCnt_B)
iResult = iCnt * 2 + iCnt_B * 2
End Function
Sub Test_2()
Dim iCnt As Integer
Dim iCnt_B As Integer
Dim iResult As Integer
iCnt = 2
iCnt_B = 3
iResult = fTest_2(iCnt, iCnt_B)
End Sub
Public Function fTest_2(iCnt, iCnt_B)
fTest_2 = iCnt * 2 + iCnt_B * 2
End Function
Let me know if not clear.
This Daily Dose of Excel conversation will be helpful.
When you use the parentheses you are forcing VBA to evaluate what's inside them and adding the result to the collection. Since NewFund has no default property - I assume - the evaluation yields nothing, so can't be added. Without the parentheses it evaluates to the instance of the class, which is what you want.
Another example. This:
Dim coll As Collection
Set coll = New Collection
coll.Add Range("A1")
Debug.Print coll(1); TypeName(coll(1))
and this ...
coll.Add (Range("A1"))
Debug.Print coll(1); TypeName(coll(1))
... both yield whatever is in A1 in the debug.window, because Value is Range's default property. However, the first will yield a type of "Range", whereas the type in the 2nd example is the data type in A1. In other words, the first adds a range to the collection, the 2nd the contents of the range.
On the other hand, this works:
Dim coll As Collection
Set coll = New Collection
coll.Add ActiveSheet
Debug.Print coll(1).Name
... and this doesn't:
coll.Add (ActiveSheet)
Debug.Print coll(1).Name
because ActiveSheet has no default property. You'll get an runtime error 438, just like in your question.
Here's another way of looking at the same thing.
Let assume that cell A1 contains the string Hi!
Function SomeFunc(item1, item2)
SomeFunc = 4
End Function
Sub Mac()
' here in both of the following two lines of code,
' item1 will be Variant/Object/Range, while item2 will be Variant/String:
SomeFunc Range("A1"), (Range("A1"))
Let i = SomeFunc(Range("A1"), (Range("A1")))
'this following is a compile syntax error
SomeFunc(Range("A1"), (Range("A1")))
' while here in both the following two lines of code,
' item1 will be Variant/String while item2 will be Variant/Object/Range:
SomeFunc ((Range("A1")), Range("A1")
Let j = SomeFunc((Range("A1")), Range("A1"))
'this following is a compile syntax error
SomeFunc((Range("A1")), Range("A1"))
Set r = Range("A1") ' sets r to Variant/Object/Range
Set r = (Range("A1")) ' runtime error 13, type mismatch; cannot SET r (as reference) to string "Hi!" -- Strings are not objects in VBA
Set r = Range("A1").Value ' runtime error (same)
Let r = Range("A1") ' set r to "Hi!" e.g. contents of A1 aka Range("A1").Value; conversion to value during let = assignment
Let r = (Range("A1")) ' set r to "Hi!" e.g. contents of A1 aka Range("A1").Value; conversion to value by extra ()'s
Let r = Range("A1").Value ' set r to "Hi!" by explicit use of .Value
End Sub
I only add this to help illustrate that there are two things going on here, which could be conflated.
The first is that the () in an expression that converts the item to its Value property as stated above in other answers.
The second is that functions invoked with intent to capture or use the return value require extra () surrounding the whole argument list, whereas functions (or sub's) invoked without intent to capture or use the return value (e.g. as statements) must be called without those same () surrounding the argument list. These surrounding () do not convert the argument list using .Value. When the argument list has only one parameter, this distinction can be particularly confusing.

Sorting a collection of objects in VBA

I'm trying to write a function that would sort a collection of objects. Since the objects are all of the same type (the same user-defined class), their property set is the same. Is it possible to discover the object's properties (through code) so as to put the collection in a bi-dimensional array, each row being for an object, each column for one of its property?
Another solution would be to copy each object from the collection to an array of objects, and sort them by one of their property, whose name is passed to the function as a string. But I don't see how I can point to the object's property using the property's name passed as a string.
For a collection, it is best to sort it by it's Keys (that's what they're there for) -- but in case you don't have the keys list (lost your keys!):
'Give an input "Data As Collection"
Dim vItm As Variant
Dim i As Long, j As Long
Dim vTemp As Variant
For i = 1 To Data.Count – 1
For j = i + 1 To Data.Count
If CompareKeys(Data(i).myMemberKey, Data(j).myMemberKey) Then
'store the lesser item
vTemp = Data(j)
'remove the lesser item
Data.Remove j
're-add the lesser item before the greater Item
Data.Add vTemp, , i
End If
Next j
Next i
Come up with your own CompareKey function which will return true or false if the UDT member variables are >, < or 0 to one another. The reason you have to delete and re-add is because you cannot 'swap' internal members in a vb6/vba collection object.
Best of luck
EDIT:
To access a property you have the name of programmatically (as a string), use VB's CallByName function in the form:
Result = CallByName(MyObject, "MyProperty", vbGet)

How to check whether a variant array is unallocated?

Dim Result() As Variant
In my watch window, this appears as
Expression | Value | Type
Result | | Variant/Variant()
How do I check the following:
if Result is nothing then
or
if Result is Not Set then
This is basically what I am trying to accomplish, but the first one does not work and the second does not exist.
To avoid error handling, I used this, seen on a forum long time ago and used sucessfully since then:
If (Not Not Result) <> 0 Then 'Means it is allocated
or alternatively
If (Not Not Result) = 0 Then 'Means it is not allocated
I used this mainly to extend array size from unset array this way
'Declare array
Dim arrIndex() As Variant
'Extend array
If (Not Not Result) = 0 Then
ReDim Preserve Result(0 To 0)
Else
ReDim Preserve Result(0 To UBound(Result) + 1)
End If
Chip Pearson made a useful module called modArraySupport that contains a bunch of functions to test for things like this. In your case, you would want to use IsArrayAllocated.
Public Function IsArrayAllocated(Arr As Variant) As Boolean
This function returns TRUE or FALSE indicating whether the specified array is allocated (not empty). Returns TRUE of the
array is a static array or a dynamic that has been allocated with a Redim statement. Returns FALSE if the array is a dynamic array that
has not yet been sized with ReDim or that has been deallocated with the Erase statement. This function is basically the opposite of
ArrayIsEmpty. For example,
Dim V() As Variant
Dim R As Boolean
R = IsArrayAllocated(V) ' returns false
ReDim V(1 To 10)
R = IsArrayAllocated(V) ' returns true
The technique used is basically to test the array bounds (as suggested by #Tim Williams) BUT with an extra gotcha.
To test in your immediate window:
?IsArrayAllocated(Result)
Testing in Watch window: there are may ways to do this; for example, add a watch on R and under "Watch Type" select "Break When Value Changes".
You can use the following in the immediate window:
?Result Is Nothing
?IsNull( Result )
?IsEmpty( Result )
?IsMissing( Result )
The first is simply for completeness. Since Result is not an object, Result Is Nothing will throw an error. Empty is for variants that have not been initialized including arrays which have not been dimensioned..
(Update) In doing some additional checking, I have discovered that IsEmpty will never return true on a declared array (whether Redim'd or not) with only one exception. The only exception I found is when the array is declared at the module level and not as Public and then only when you check it in the immediate window.
Missing if for optional values passed to a function or sub. While you cannot declare Optional Foo() As Variant, you could have something like ParamArray Foo() As Variant in which case if nothing is passed, IsMissing would return true.
Thus, the only way to determine if the array is initialized is to write a procedure that would check:
Public Function IsDimensioned(vValue As Variant) As Boolean
On Error Resume Next
If Not IsArray(vValue) Then Exit Function
Dim i As Integer
i = UBound(Bar)
IsDimensioned = Err.Number = 0
End Function
Btw, it should be noted that this routine (or the library posted by Jean-François Corbett) will return false if the array is dimensioned and then erased.
I recommend a slightly different approach because I think using language artifacts like (Not Array) = -1 to check for initialization is difficult to read and causes maintenance headaches.
If you are needing to check for array allocation, most likely it's because you're trying to make your own "vector" type: an array that grows during runtime to accommodate data as it is being added. VBA makes it fairly easy to implement a vector type, if you take advantage of the type system.
Type Vector
VectorData() As Variant
VectorCount As Long
End Type
Dim MyData As Vector
Sub AddData(NewData As Variant)
With MyData
' If .VectorData hasn't been allocated yet, allocate it with an
' initial size of 16 elements.
If .VectorCount = 0 Then ReDim .VectorData(1 To 16)
.VectorCount = .VectorCount + 1
' If there is not enough storage for the new element, double the
' storage of the vector.
If .VectorCount > UBound(.VectorData) Then
ReDim Preserve .VectorData(1 To UBound(.VectorData) * 2)
End If
.VectorData(.VectorCount) = NewData
End With
End Sub
' Example of looping through the vector:
For I = 1 To MyData.VectorCount
' Process MyData.VectorData(I)
Next
Notice how there's no need to check for array allocation in this code, because we can just check the VectorCount variable. If it's 0, we know that nothing has been added to the vector yet and therefore the array is unallocated.
Not only is this code simple and straightforward, vectors also have all the performance advantages of an array, and the amortized cost for adding elements is actually O(1), which is very efficient. The only tradeoff is that, due to how the storage is doubled every time the vector runs out of space, in the worst case 50% of the vector's storage is wasted.
Check the LBound of the array. If you get an error then it's uninitialized.

Passing and Receiving an Array from Function

I have made a bunch of 2D arrays in Excel, and I have written a function to put the same data in a copy of each. I'm not quite sure I'm getting the fiddly parts of the syntax correct.
The function is called 'Fill', the old array 'Old' and the new one 'New'. I'm using the name 'Block' for the passing-between variable name.
So, the line in my code is:
New = Fill(Block:=Old())
And the first line of my function is:
Function Fill(Block() As Variant) As Variant
This gives me a type mismatch error of the 'Old' array, saying it is expecting an array. Leads me to think the function is OK, and waiting for the correct type, but not receiving it.
What am I missing?
It has been a while since I did VBA programming, but I would think the following is more likely to be correct:
NewArray = Fill(OldArray)
Function Fill(Block As Variant) As Variant
Fill = Block
End Function
Here are some notes on why you may have got the error you did. If a function expects a particular type, you must declare the variable as that type.
Sub FillThis()
'Declare OldArray as Variant '
'because that is what the function '
'requires. '
Dim OldArray As Variant
'Similarly ...'
Dim StringPart As String
'ByRef, so the variable will be '
'changed by the function. '
'Note that ByRef is the default. '
Fill OldArray
For i = 0 To 4
Debug.Print OldArray(i)
Next
StringPart = "Part 1"
GetString StringPart
Debug.Print StringPart
End Sub
'In this example, Fill is not being returned, '
'so there is no need to declare it as anything '
Function Fill(ByRef OldArray As Variant)
'The Array must be dimensioned '
ReDim OldArray(4)
For i = 0 To 4
OldArray(i) = i + 1
Next
End Function
Function GetString(ByRef StringPart As String)
StringPart = StringPart & " Add a Bit"
End Function