BINARY SEARCH:

1) Declare your definition of what a binary search is.

2) Define a table with 10 sorted entries (elements) which your program will do the search on.

3) Write and post the code to do a binary search on the above table, searching for a user provided argument, and issuing a found or not found message at the end.

*****

Huh?

a binary search goes through an array by first taking its size divided by 2, then only asking if the number is higher or lower than the number at the current entry. After knowing if it's higher or lower, it halves itself in the according direction.

I couldn't figure out how to code this (I'm stupid) and I'm not sure how accurate my description is.

Code:

`DECLARE FUNCTION binarySearchRec% (Array%(), i1%, i2%, n%)`

DECLARE FUNCTION binarySearch% (Array%(), n%)

' Binary search example by Na Than

CLS : RANDOMIZE TIMER

' First make a sorted array

DIM Array%(15)

PRINT "POSTN:"

PRINT "CONT:"

FOR i% = 0 TO 15

Array%(i%) = a%

LOCATE 1, 8 + i% * 3: PRINT i%

LOCATE 2, 8 + i% * 3: PRINT a%

a% = a% + INT(RND * 5)

NEXT i%

PRINT

' Test loop. Enter sumthin <0 to exit

DO

PRINT "Find what";

INPUT n%

IF n% < 0 THEN EXIT DO

Res% = binarySearch%(Array%(), n%)

IF Res% <> -1 THEN

PRINT "Found at "; Res%

ELSE

PRINT "Not Found"

END IF

LOOP

'

' This function searches for n% in Array%() returning its position,

' or "-1" if n% is not in Array%().

'

FUNCTION binarySearch% (Array%(), n%)

binarySearch% = binarySearchRec%(Array%(), LBOUND(Array%), UBOUND(Array%), n%)

END FUNCTION

'

' binarySearchRec% is called by binarySearch% takes four parameters:

'

' Array%() -> The array to look inside of.

' i1% -> lower limit.

' i2% -> upper limit.

' n% -> Number that is searched.

'

' We basicly take the array, and look the element in the mid place.

' if it equals n%, we have found it!

'

' If the element in the array is SMALLER than n%, we should look in

' the right half of the array, so we call ourselves recursively with

' new i1%, i2% values.

'

' If the element in the array is BIGGER than n%, we should look in the

' left half of the array.

'

' This (of course) only works if the array is sorted.

'

FUNCTION binarySearchRec% (Array%(), i1%, i2%, n%)

Res% = -1

IF i1% > i2% THEN binarySearchRec% = Res%: EXIT FUNCTION

midPoint% = i1% + (i2% - i1%) \ 2

IF Array%(midPoint%) = n% THEN

Res% = midPoint%

ELSEIF Array%(midPoint%) < n% THEN

Res% = binarySearchRec%(Array%(), midPoint% + 1, i2%, n%)

ELSE

Res% = binarySearchRec%(Array%(), i1%, midPoint% - 1, n%)

END IF

binarySearchRec% = Res%

END FUNCTION

Can anyone point to a DL of Address.bas? It has a binary search algo in there. ;*)

*curses CT....

WHITETIGER:

---------------

See Ninkazu's definition which is pretty good.

An example of a binary search that you may have seen is:

You ask a person to think of a number from 1 to 100. You can find his number in 7 questions. Why? Because 2 to the 7th power is 128 which is greater than 100. Each of your questions is going to eliminate half of the possibiities. You start your questions at the mid-point of 50. You first ask if the answer is less than 50. If yes, your next question will be against 25. If no, your next question will be against 75, and so on.

NINKAZU:

-----------

I think you have the idea. Why don't you try it.

NATHAN:

----------

I certainly did not expect a recursive solution. I honestly don't think recursion is required for this. It only takes about 16 lines of code to do it straightforward. I have not checked it out yet because the PC that I'm using right now doesn't have my compiler. Later.

*****

NATHAN,

Wonderful! Your program works vey well --- a very complete solution. It never would have occurred to me to use recursion. You should post this program in Antoni's recursion challenge. It's a very strong candidate to be the winner.

*****

Thanks man. In fact, I did it very naturally. Sometimes, the recursive sollution is the easy one for me, as I tend to think recursively in most cases.

And, sinceramente, no se me ocurre cÃ³mo hacerlo de forma iterativa

Nathan,

Don't forget to post this under Antoni's topic.

*****

A binary search checks the middle item of those that have not yet been eliminated, and (depending on whether the item is too high or too low) eliminates either the top half or the bottom half from consideration. It does this over and over until it has either found the item or exhausted all possibilites.

Code:

`DECLARE FUNCTION binarySearch% (item%, first%, last%, array%(), where%)`

DEFINT A-Z

'---------------------------------------

' Define a random sorted array:

RANDOMIZE TIMER

DIM array(1 TO 10)

n = 0

FOR i = 1 TO 10

n = n + INT(RND * 100)

array(i) = n

NEXT i

'---------------------------------------

'---------------------------------------

' Display the elements of the array:

CLS

PRINT

PRINT "Items in the array:"

FOR i = 1 TO 10

PRINT array(i);

NEXT i

PRINT

'---------------------------------------

'---------------------------------------

' Ask the user for an item to search for:

PRINT

INPUT "Item to search for: ", item

PRINT

'---------------------------------------

'---------------------------------------

' Attempt to find it, then display the result:

IF binarySearch(item, 1, 10, array(), position) THEN

PRINT item; "was found at position"; position

ELSE

PRINT item; "was not found in the list."

END IF

'---------------------------------------

END

'

' Returns false if the item isn't found, or true if it was in which

' case where% is the item's index.

'

FUNCTION binarySearch (item, first, last, array(), where)

DO WHILE first <= last

middle = (first + last) \ 2

i = array(middle)

IF i = item THEN binarySearch = -1: where = middle: EXIT FUNCTION

IF i < item THEN first = middle + 1 ELSE last = middle - 1

LOOP

binarySearch = 0

END FUNCTION