Looping through array, deleting elements

My idea of looping through an array, interrogating elements that are
of class instruction, and deleting the element does not appear to work
as I thought. I think I know why. What is the best solution?

block.each {|i| ab.delete(i) if i.parm == "FOO"}

where block is a series of elements of class instruction(i). Class
instruction has two attributes, parm and arg. If the instruction parm
is FOO, then delete the instruction object from the array and keep
looping. Half of what should be deleted actually get's deleted. I
think the reason is: the loop is looking at element x foo, element x
foo get's deleted, the array is shifted left, array pointer has not
changed but is now referencing element y foo, loop advances the array
pointer, and now array pointer is pointing at element z foo.

If what I think is correct, what is the best solution? Do I set the
current element to nil, then compact the array after the loop? Do I
delete the instruction and then do a redo? Or is there a better
solution than those?

Thanks,
dvn

dkmd_nielsen wrote:

My idea of looping through an array, interrogating elements that are
of class instruction, and deleting the element does not appear to work
as I thought. I think I know why. What is the best solution?

What about the Array#delete_if method?

···

--
Posted via http://www.ruby-forum.com/.

This sounds like it would be a lot easier with Array#reject or Array#select.

···

On 08/03/07, dkmd_nielsen <donn@cmscms.com> wrote:

My idea of looping through an array, interrogating elements that are
of class instruction, and deleting the element does not appear to work
as I thought. I think I know why. What is the best solution?

block.each {|i| ab.delete(i) if i.parm == "FOO"}

where block is a series of elements of class instruction(i). Class
instruction has two attributes, parm and arg. If the instruction parm
is FOO, then delete the instruction object from the array and keep
looping. Half of what should be deleted actually get's deleted. I
think the reason is: the loop is looking at element x foo, element x
foo get's deleted, the array is shifted left, array pointer has not
changed but is now referencing element y foo, loop advances the array
pointer, and now array pointer is pointing at element z foo.

If what I think is correct, what is the best solution? Do I set the
current element to nil, then compact the array after the loop? Do I
delete the instruction and then do a redo? Or is there a better
solution than those?

Thanks,
dvn

Alle giovedì 8 marzo 2007, dkmd_nielsen ha scritto:

My idea of looping through an array, interrogating elements that are
of class instruction, and deleting the element does not appear to work
as I thought. I think I know why. What is the best solution?

block.each {|i| ab.delete(i) if i.parm == "FOO"}

where block is a series of elements of class instruction(i). Class
instruction has two attributes, parm and arg. If the instruction parm
is FOO, then delete the instruction object from the array and keep
looping. Half of what should be deleted actually get's deleted. I
think the reason is: the loop is looking at element x foo, element x
foo get's deleted, the array is shifted left, array pointer has not
changed but is now referencing element y foo, loop advances the array
pointer, and now array pointer is pointing at element z foo.

If what I think is correct, what is the best solution? Do I set the
current element to nil, then compact the array after the loop? Do I
delete the instruction and then do a redo? Or is there a better
solution than those?

Thanks,
dvn

I'm not sure I understand your code (what is that ab in the each block?) At
any rate, there's method which directly deletes the elements of an array
which satisfy a given condition: Array#delete_if:

a=[1,2,3,4,5,6]
a.delete_if{|i| i > 3}
=> [1,2,3]

This method passes each element of the array to the block, then deletes it if
the block returns true.

Stefano

My idea of looping through an array, interrogating elements that are
of class instruction, and deleting the element does not appear to work
as I thought. I think I know why. What is the best solution?

block.each {|i| ab.delete(i) if i.parm == "FOO"}

Not a good idea! You are modifying the array as you iterate over it
-- this will cause some unexpected behavior:

ary = [1,2,3,4]
ary.each {|x| ary.delete(x) if x > 2}

p ary #=> [1, 2, 4]

oops!

where block is a series of elements of class instruction(i). Class
instruction has two attributes, parm and arg. If the instruction parm
is FOO, then delete the instruction object from the array and keep
looping. Half of what should be deleted actually get's deleted. I
think the reason is: the loop is looking at element x foo, element x
foo get's deleted, the array is shifted left, array pointer has not
changed but is now referencing element y foo, loop advances the array
pointer, and now array pointer is pointing at element z foo.

You are correct. What happens is that we delete the number 3 from the
array and this compacts the array -- 4 shifts into the position where
3 just was. Now the "each" method moves the index to the next position
thereby skipping the number 4.

If what I think is correct, what is the best solution? Do I set the
current element to nil, then compact the array after the loop? Do I
delete the instruction and then do a redo? Or is there a better
solution than those?

Use the delete_if method on the array.

ary.delete_if {|x| x > 2}

p ary #=> [1,2]

Blessings,
TwP

···

On 3/8/07, dkmd_nielsen <donn@cmscms.com> wrote:

There are a couple of dumb things I have done. First, I had omitted
that block is an object, not of class array. So its each method is
iterating through an array internal to it. Nice little detail to
omit. Duh. Second, there is something called "retry". I was aware
of break, next, and redo, but I was unware of "retry." It worked
great.

I looked at the delete_if. What I could do with that is create a
delete_if method in the class block that implements delete_if. In the
long run, I think that is the best thing to do.

Thanks for everything. One always finds the answer on his/her own
shortly after asking for help.

dvn

···

On Mar 8, 3:02 pm, "Tim Pease" <tim.pe...@gmail.com> wrote:

On 3/8/07, dkmd_nielsen <d...@cmscms.com> wrote:

> My idea of looping through an array, interrogating elements that are
> of class instruction, and deleting the element does not appear to work
> as I thought. I think I know why. What is the best solution?

> block.each {|i| ab.delete(i) if i.parm == "FOO"}

Not a good idea! You are modifying the array as you iterate over it
-- this will cause some unexpected behavior:

ary = [1,2,3,4]
ary.each {|x| ary.delete(x) if x > 2}

p ary #=> [1, 2, 4]

oops!

> where block is a series of elements of class instruction(i). Class
> instruction has two attributes, parm and arg. If the instruction parm
> is FOO, then delete the instruction object from the array and keep
> looping. Half of what should be deleted actually get's deleted. I
> think the reason is: the loop is looking at element x foo, element x
> foo get's deleted, the array is shifted left, array pointer has not
> changed but is now referencing element y foo, loop advances the array
> pointer, and now array pointer is pointing at element z foo.

You are correct. What happens is that we delete the number 3 from the
array and this compacts the array -- 4 shifts into the position where
3 just was. Now the "each" method moves the index to the next position
thereby skipping the number 4.

> If what I think is correct, what is the best solution? Do I set the
> current element to nil, then compact the array after the loop? Do I
> delete the instruction and then do a redo? Or is there a better
> solution than those?

Use the delete_if method on the array.

ary.delete_if {|x| x > 2}

p ary #=> [1,2]

Blessings,
TwP