Like your previous question, both of these are misread answers from gimmick crosswords.

DRAWI is from the New York Times crossword, March 4 2009. I can't see the actual crossword since I don't have a subscription, but from this link, we can see the puzzle's gimmick. Several squares could be filled with "in" or "out", and these squares would be doubly clued. For instance, DRAW[IN/OUT] was clued as "Attract / Protract".

TAKEF is also from the NYT crossword, this time from August 17, 2003. This puzzle's gimmick was that answers having numbers were entered so that the number fit in one box. For instance, "zero feet" was entered as [ZERO]FEET.

In both cases, the parser was incorrectly reading only the first letter of each rebus square. I suggest attempting to fill the grid yourself, downloading an alternate dictionary for your crossword filler, or using software such as QXW that will let you choose words to eliminate nonsense answers such as these.

Like your previous question, both of these are misread answers from gimmick crosswords.

DRAWI is from the New York Times crossword, March 4 2009. I can't see the actual crossword since I don't have a subscription, but from this link, we can see the puzzle's gimmick. Several squares could be filled with "in" or "out", and these squares would be doubly clued. For instance, DRAW[IN/OUT] was clued as "Attract / Protract".

TAKEF is also from the NYT crossword, this time from August 17, 2003. This puzzle's gimmick was that answers having numbers were entered so that the number fit in one box. For instance, "zero feet" was entered as [ZERO]FEET.

In both cases, the parser was incorrectly reading only the first letter of each rebus square. I suggest attempting to fill the grid yourself, downloading an alternate dictionary for your crossword filler, or using software such as QXW that will let you choose words to eliminate nonsense answers such as these.