View this file with results and syntax highlighting here.
Find
Find (β·) searches for occurrences of an array π¨ within π©. The result contains a boolean for each possible location, which is 1 if π¨ was found there and 0 if not.
"xx" β· "xxbdxxxcx"
More precisely, π¨ needs to match a contiguous selection from π©, which for strings means a substring. These subarrays of π© are also exactly the cells in the result of Windows. So we can use Windows to see all the arrays π¨ will be compared against.
2 β "xxbdxxxcx"
"xx"βΈβ‘Λ 2 β "xxbdxxxcx"
Like Windows, the result usually doesn't have the same dimensions as π©. This is easier to see when π¨ is longer. It differs from APL's version, which includes trailing 0s in order to maintain the same length. Bringing the size up to that of π© is easy enough with Take (β), while shortening a padded result would be harder.
"string" β· "substring"
"string" (β’ββ’ββ·) "substring" # APL style
If π¨ is larger than π©, the result is empty, and there's no error even in cases where Windows would fail. One place this tends to come up is when applying First (β) to the result: ββ· tests whether π¨ appears in π© at the first position, that is, whether it's a prefix of π©. If π¨ is longer than π© it shouldn't be a prefix. First will fail but using a fold 0β£Β΄β· instead gives a 0 in this case.
"loooooong" β· "short"
9 β "short"
0 β£Β΄ "loooooong" β· "short"
Adding a Deshape gives 0β£Β΄β₯ββ·, which works with the high-rank case discussed below. It tests whether π¨ is a multi-dimensional prefix starting at the lowest-index corner of π©.
Higher ranks
If π¨ and π© are two-dimensional then Find does a two-dimensional search. The cells used are also found in π¨β’βΈβπ©. For example, the bottom-right corner of π© below matches π¨, so there's a 1 in the bottom-right corner of the result.
β’ a β 7 (4|βΛ)βββ 9 # Array with patterns
(0βΏ3βΏ0β0βΏ1βΏ0) β· a
It's also allowed for π¨ to have a smaller rank than π©; the axes of π¨ then correspond to trailing axes of π©, so that leading axes of π© are mapped over. This is a minor violation of the leading axis principle, which would match axes of π¨ to leading axes of π© in order to make a function that's useful with the Rank operator, but such a function would be quite strange and hardly ever useful.
0βΏ1βΏ0βΏ1 β· a