Page 1 of 1
I need a sorting algorythm
Posted: Wed Nov 08, 2006 2:17 am
by bigbluehacker
I need an algorythm to sort one field of an array while preserving the relationship with all other fields on the same line. For example:
1 Man-O-War $3,967,845
2 Swaps $2,897,136
3 Secretariat $10,145,873
4 Sunday Silence $265,972
These are fictional examples; I need to sort winnings by number; don't need the name. result would look like:
4 265,972
2 2,897,136
1 3,967,845
3 10,145,873
Years ago I had such a routine written in Commodore Basic and have long since lost it. It was about 20 lines long. I haven't been able to find such a routine for QBasic and have searched everywhere. Does anyone know of one?
Posted: Sat Dec 30, 2006 6:41 am
by Dr_Davenstein
Hmmm... Do you mean you have different fields stored in a linear fashion within the same array?
Anyway, the easiest type of sort to learn, is the bubble sort. It's basically like this...
Code: Select all
Dim Array(20) As Integer
For i = LBound(Array) To UBound(Array)
y=y+1
Array(i) = Int(Rnd*1000)
Print Array(i)
Next
For i = LBound(Array) To UBound(Array)
For j = i To UBound(Array)
If Array(i)>=Array(j) Then
Swap Array(i), Array(j)
End If
Next
Next
y=0
For i = LBound(Array) To UBound(Array)
y=y+1
Locate y,10
Print Array(i)
Next
Posted: Sat Jan 06, 2007 12:00 am
by bigbluehacker
Thank you for your reply.
I have tried to use "bubble" and other simple sorts and can't get them to work correctly with 2 or 3 dimensional arrays. If DIM Array(1 to 20,1 to 20) AS LONG I can sort elements in one of the dimensions, but the other dimension remains unaltered.
If I dimension the array as Sales(1 to 200, 1 to 200) as string (because the first dimension is names and the second is numbers), how can I sort the sales amounts from high to low and read the correct salesperson's name next to each amount?
Thanks again for your input. Any other suggestions will be welcome.
Posted: Sat Jan 06, 2007 12:48 am
by Ralph
How about concatinating field 3 with a spacer and with field 1? Then, you can sort the new, concatenated values, after which you would again separate the components, by using the LEFT$, MID$, AND RIGHT$, as covenient.
Do you see the logic?
Posted: Sat Jan 06, 2007 5:48 am
by buff1
Set up 3 arrays.
test data.
1 Man-O-War $3,967,845
2 Swaps $2,897,136
3 Secretariat $10,145,873
4 Sunday Silence $265,972
array1$(1)="1":array1$(2)="2" etc
array2$(1)="Man-O-War":array2$(2)="Swaps" etc.
array3$(1)="3967845":array3$(2)="2987136" etc.
then sort the 'array' that you want sorted and use swap to swap the
other array with the same item numbers
i.e. if array1$ is used for the sort
then when you
swap array1$(j),array1$(k)
you also
swap array2$(j),array2$(k)
swap array3$(j),array3$(k)
Posted: Sat Jan 06, 2007 9:31 pm
by Ralph
I have to admit that Buff1's method is very good! Forget my suggestion.
Posted: Fri Jan 19, 2007 8:52 am
by Dr_Davenstein
Does QB allow you to swap whole UDT's of the same TYPE? I can't remember. That would be the best approach, in my opinion. If that doesn't work, I would definetly go with buff1.
