Need code to load Treeview from file path list

[Paul Squires]

Hi Everyone,

I'm trying not to be lazy but I'll ask, does anyone have some code that they're willing to share that will take an array of file paths and load them into a Treeview so that the Treeview layout is equal to the directory structure.

I have a string array of file paths (sorted and capitalized) and I need to walk the array from the first element to the last and create the correct nodes in the correct relationship (parent/child).

.... just trying to save myself a few hours of headache trying to figure out this algorithm.

Thanks!

[George Bleck]

Try this, might be able to modify it to suit your needs.

http://www.powerbasic.com/support/pbforums/showthread.php?t=23778

[Paul Squires]

Thanks George - I'll take a look.

[Michael Mattias]

GIVEN:

S() sorted array which looks like...

C:
C:\DOG
C:\DOG\CAT
C:\DOG\CAT2
D:
D:\DOG
D:\DOG\CAT
D:\DOG\CAT2

LOCAL hParentITem () AS LONG 

REDIM hParentItem(50)   ' enough to hold every possible subfolder  

FOR Z = LBOUND (S,1) TO UBOUND (S,1)  ' for each Drive/Folder in S() 

  ' is it a drive? 
    IF RIGHT$(S(Z) = ":" THEN 
      iLevel =  0               ' root 
    ELSE
     ' find the 'depth' or 'level' of this node: 
       iLevel = TALLY(s(Z), "\")
    END IF 
    hParentITem(iLevel) =  TV_INSERTITEM (Parent= IIF&(iLevel=0, %TV_ROOT, hParentItem(iLevel-1))) 
    END IF 
NEXT

Needs some tweaking but it's a start. The key is keeping track of the last "hitem" by "level" ("depth") and the array thing has always worked for me.

MCM

[Paul Squires]

Thanks guys,

I was able to build some code that works really well. The core code that does the work is shown below. Granted, it is pretty application specific so you won't be able to simply cut and paste it into an application.

The general idea is to iterate through the sorted array list. Whenever the path of the string changes then we need to create a new parent node. The trick is to ensure that each piece of the path already has a parent node that exists. To do that check I look up the piece of the path in a Hash Table. The hash table entry holds the peice of the path and its associated treeview node handle. Holding the data in a hash table makes processing the strings extremely fast. I have posted the hash table code elsewhere in these forums.

   Array Sort sFiles(), TagArray nIndex(), Collate UCase

   ' Create the hash table
   pDict = hash_create( 100, %FALSE )
   
   For y = 1 To nCount
      ' If the text has changed then we need to add a new parent node.
      If ( y = 1) Or (y > 1 And sFiles( y - 1 ) <> sFiles( y )) Then
         ' Determine under which node to put this node. We start by defaulting to
         ' placing it under the root node (hNodeFunction). 
         hParentNode = hNodeFunction 

         ' Walk the string ensuring that each section of the path 
         ' exists in the Treeview (i.e. it should have the node value
         ' saved in the hash and is returned via the zData parameter).
         NumSections = ParseCount( sFiles(y), "\" )
         For z = 1 To NumSections
             st = Parse$(sFiles(y), "\", z)
             ' If the section exists in the Hash then update the hParentNode
             ' that we will use when adding the path to the Treeview.
             If Hash_Find( pDict, st, zData ) Then
                hParentNode = zData
             Else
                hParentNode = FF_TreeView_InsertItem( tv, hParentNode, st, %NODE_CATEGORY, _  
                                                          g.Explorer_ClosedIcon, g.Explorer_OpenIcon )   
                Hash_Add pDict, st, hParentNode
             End If
         Next
         
      End If
      
      CodeItem = gclsCodestore.GetCodeItem(nIndex(y))
      FF_TreeView_InsertItem tv, hParentNode, CodeItem.FunctionName, 0, _
                                  g.Explorer_CodeIcon, g.Explorer_CodeIcon    
   Next
   
   hash_destroy pDict

[Michael Mattias]

Your Key/Data approach is imaginative.

I like imaginative.

[jcfuller]

I found that using IDictionary is faster than Paul's hash table. I was quite surprised.

There is a test piece attached to one of my replies in this thread.

http://www.powerbasic.com/support/pbforums/showthread.php?t=38402&page=2

James