https://bugs.documentfoundation.org/show_bug.cgi?id=154125

--- Comment #3 from ady <[email protected]> ---
Using LO 7.4.5, here is what I have tested ("|" is column separator, ";" as
both row and argument separators).

In a pre-dynamic formula context, with INDEX() in array form *by itself*:


A_
{ =INDEX({11|12;21|22};0;1) } (with CSE)
=11 (1st row, 1st col)
=21 (2nd row, 1st col)

OK.


B_
=INDEX({11|12;21|22};0;1) (_no_ CSE)
=11 (1st row only)

OK.


C_
{ =INDEX({11;21};0;1) } (with CSE)
=11 (1st row)

NOT OK. We should also get:
=21 (2nd row)

D_
=INDEX({11;21};0;1) (_no_ CSE)
=11 (1st row)

OK, but expected considering result "C" above.


E_
{ =INDEX({11|12};1;0) } (with CSE)
=11 (1st col)
NOT OK. We should also get:
=12 (2nd col)

F_
=INDEX({11|12};1;0) (_no_ CSE)
=11 (1st col)

OK, but expected considering result "E" above.


Considering that those are wrong results already, I have not tested additional
possibilities such as default values when arguments are omitted, or
out-of-scope arguments, or wrong-type arguments.


Please note that these are examples of INDEX() by itself. When INDEX() is used
as lookup_array for MATCH() (as in the example in comment 0), then CSE should
not be required in order to obtain the array to be fed for MATCH().

Once the above examples get to work, I should be able to test them in
combination with MATCH() as in the example in comment 0.

PS: FWIW, in Excel I believe that =INDEX({1,2};2) should had been #REF!, IIUC
(whether with CSE or not).

-- 
You are receiving this mail because:
You are the assignee for the bug.

Reply via email to