How to remove seq of array elments matching a pattern at the beginning and the end?

Suppose that I have a json file, in which the following pattern appears many times.

... [ ... ["X"], ... ,["Y"] ... ] ...

I want to remove everything between each pair of ["X"] and ["Y"]. How can I do it?

The question seems to have two components:

(1) Given an array, how can I delete all segments that are "bookended" by two values?

(2) Given a single JSON entity (aka document), how can I perform the above-mentioned deletion operation on all arrays, no matter where they occur within the document?

Here, I will offer an alternative to @pmf's solution for (1) and show how to apply it to an entire JSON entity.

Here's the alternative, which has the possible advantage that it doesn't make any strong assumptions about the occurrence of the $x and $y values, and allows for both interpretations regarding the removal of the bookends themselves:

# Input: an array
# Remove all stretches from $x to the next $y,
# removing both bookends too if and only if $bookends.
# Both bookends must be present for a stretch to be removed.
def remove_all_xy($x; $y; $bookends):
  # The helper function removes a single stretch from $x to $y, if any
  def r:
    index($x) as $ix
    | if $ix then .[$ix+1:] as $tail
      | ($tail | index($y)) as $iy
      | if $iy  
        then (if $bookends then 0 else 1 end) as $adjust
        | .[:$ix + $adjust] + ($tail | .[1+$iy - $adjust:] | r)
        else . end
      else . end;
  r;

Now let's say you decide on some function, foo($x;$y;$bookends), for performing the per-array operation. To apply it to the whole document,
you could write:

walk(if type == "array" then foo($x;$y;$bookends) else . end)

This might not be as efficient as possible, but in practice it should suffice. (If not, then simply adapt the standard walk.)

Assuming pairwise occurrences, and that "between" means excluding the border items, you could query the indices of the border items, and fetch everything in between.

Given the following sample, the following filter will produce the following output:

[
  ["A"], ["B"], ["X"], ["C"], ["D"], ["Y"], ["E"], ["F"],
  ["G"], ["H"], ["X"], ["I"], ["J"], ["Y"], ["K"], ["L"]
]

[
  (
    [[indices([["X"]]), indices([["Y"]])] | 0, transpose[][], infinite]
    | _nwise(2)
  )
  as [$a, $b] | .[$a:$b+1][]
]

[["A"],["B"],["X"],["Y"],["E"],["F"],["G"],["H"],["X"],["Y"],["K"],["L"]]

If the border items may occur in any order, and not necessarily in equal amounts, before transposing the two lists of indices they would need to be filtered first to only contain mutually successive positions.