Yes, there are a number of possibilties
You have to find the balance between ease of use and performance
Personally, I would prefer functions a la SetItem(value, x, y, z, ..) and getItem(@value, x, y, z, ...) to avoid the access and assign methods
I suppose the access and assign won't allow more than 2 dimensions
For a couple of dimensions ( say up to 8 or so), SetItem() and GetItem() may be the fastest
If we start dreaming and go beyond 24 dimensions - we'll need an array - SetItem(value, @dimensions) with the overhead of setting the elements of dimensions before each call
Using a cursor - I am in favour of one compound index since I think a locate will be slower. But locate has the advantage you do not need to build the index expression
As to the number of dimensions I cannot imagine when/where we would need 200
All in all - I would end up using the SafeArray api
- need to build the index array - but is just string concatenation
- no need for cursors
- start and end index possible per dimension - you can have zero based arrays
- it gets a bit more challenging when you want to store strings
_
>You can fake a direct access and then implement a cursor like Al suggested, I use a PK so I can convert the subscripts (any number of dimensions up to 253) to a record (I might want to create an index on the value to have the "array" sorted) , so you end up with something like this
>
>(very crude, just spent 15 minutes on it as a proof of concept :))
>
>
>clear
>loMyClass=Createobject('myArray')
>loMyClass.setDimensions(4, 'C(16)')
>
>loMyClass.Data['1,1,1,1'] = '[001,001,001,001]'
>loMyClass.Data['4,2,1,3'] = '[004,002,001,003]'
>
>? loMyClass.Data['1,1,1,1']
>? loMyClass.Data['4,2,1,3']
>
>loMyClass.Data['4,2,1,3'] = '[EMPTY]'
>? loMyClass.Data['4,2,1,3']
>
>
>define class myArray as session
>
> DIMENSION Data(1)
>
> procedure Data_assign(vNewVal, m.nIndex1)
> LOCAL lnItem
> lnItem = this.DimensionsToItem(m.nIndex1)
> IF INDEXSEEK(PK, .t., 'cData', 'PK')
> replace Value WITH m.vNewVal IN cData
> ELSE
> this.newRow(m.nIndex1, m.vNewVal)
> ENDIF
> endfunc
>
> procedure Data_Access(m.nIndex1)
> LOCAL lnItem
> lnItem = this.DimensionsToItem(m.nIndex1)
> RETURN IIF(INDEXSEEK(PK, .t., 'cData', 'PK'), cData.Value, null)
> endfunc
>
> PROCEDURE newRow(tcDimensions as String, txValue as Variant) as VOID
>
> LOCAL laDimensions(1), lnDimensions, lnDimension
>
> lnDimensions = ALINES(laDimensions, tcDimensions, 1, ',')
> lcCommand = 'INSERT INTO cData ('
> FOR lnDimension = 1 TO lnDimensions
> lcCommand = lcCommand + IIF(lnDimension = 1, '', ', ') + 'C' + TRANSFORM(lnDimension, '@L 999')
> NEXT lnDimension
> lcCommand = lcCommand + + ', value) values ('
> FOR lnDimension = 1 TO lnDimensions
> lcCommand = lcCommand + IIF(lnDimension = 1, '', ', ') + TRANSFORM(laDimensions[lnDimension])
> NEXT lnDimension
> lcCommand = lcCommand + ', ' + TRANSFORM(txValue) + ')'
> EXECSCRIPT(lcCommand)
>
> RETURN null
> ENDPROC
>
> PROCEDURE setDimensions(tnDimensions as Integer, tcValueType as String) as VOID
> LOCAL lcCommand, lnDimension, lnDimension
>
> lcCommand = 'CREATE CURSOR cData (PK I AutoInc'
> FOR lnDimension = 1 TO tnDimensions
> lcCommand = lcCommand + ', C' + TRANSFORM(lnDimension, '@L 999') + ' I'
> NEXT lnDimension
> lcCommand = lcCommand + ', Value ' + tcValueType + ')'
> EXECSCRIPT(lcCommand)
> INDEX on PK TAG PK
> * Index on every column to speed up access?
> RETURN NULL
> ENDFUNC
>
> FUNCTION DimensionsToItem(tcDimensions as String) as Integer
>
> LOCAL laDimensions(1), lnDimensions, lnDimension
>
> lnDimensions = ALINES(laDimensions, tcDimensions, 1, ',')
> lcCommand = 'LOCATE FOR '
> FOR lnDimension = 1 TO lnDimensions
> lcCommand = lcCommand + IIF(lnDimension = 1, '', ' and ') + 'C' + TRANSFORM(lnDimension, '@L 999') + '=' + TRANSFORM(laDimensions[lnDimension])
> NEXT lnDimension
> EXECSCRIPT(lcCommand)
> RETURN IIF(FOUND(), cData.PK, - 1)
> ENDFUNC
>enddefine
>
Gregory
