Extracting table values if values in two columns have identical values in other rows

SergeiBaklan 

Well spotted!  And I think the formula wins the time trial.  When I was timing the original formula with helper ranges the calculation was taking about 230ms.  Combining them into

= LET(
    prodDecr,    product & "|" & decrement,
    exactlyOnce, UNIQUE(prodDecr,, TRUE),
    multiple?,   ISNA(XMATCH(prodDecr, exactlyOnce)),
    FILTER(Table1, multiple?)
  )

and the time required leapt to 8000ms!  Forcing the binary search and the time reduced to 150ms.  I am so used to the idea that XLOOKUP and XMATCH perform sorts as part of the solution process to gain efficiency that this came as a surprise.

Once as I had come to terms with the fact that COUNTIFS does not perform well on datasets of this size (10,000ms), I tried a different strategy of sorting the dataset and comparing each record with prior and subsequent records to determine the filter criterion.

=LET(
    sortedTbl,     SORT(Table1, {2,3}),
    CurrentPD,     CHOOSECOLS(sortedTbl,2,3),
    priorPD,       DROP(VSTACK({0,0}, CurrentPD),-1),
    nextPD,        DROP(VSTACK(CurrentPD,{0,0}),1),
    matchesnext,   BYROW(CurrentPD = nextPD, AND),
    matchesPrior,  BYROW(CurrentPD = priorPD, AND),
    matchesEither, MAP(matchesPrior, matchesnext, OR),
    FILTER(sortedTbl, matchesEither)
  )

That was back down to 250ms so, no prizes, but a pretty credible performance!

SergeiBaklan Good tip. The [search_mode] never even crossed my mind when I was running my earlier tests (heavy sigh). This is the perfect situation to implement Binary search for improved performance. Thanks for that.