Forum Discussion
Accumulating arrays
Spreadsheets are the home of the 2D array. On closer examination most of the 2D arrays that one sees are actually 'arrays of arrays', that is a 1D array nested within a list of similar arrays. At p...
I think scan_thunks was missing a closing parenthesis
= LAMBDA(init_vals_arr,array,fn,
LET(
rows_, ROWS(array),
cols_, COLUMNS(array),
thunk, LAMBDA(x,LAMBDA(x)),
row_thunks, BYROW(array, thunk),
scan_thunks,
MAP(init_vals_arr, row_thunks,
LAMBDA(init,row_thunk,
thunk(SCAN(init,row_thunk(),fn)))),
MAKEARRAY(rows_,cols_,
LAMBDA(i,j,
INDEX( INDEX(scan_thunks,i,1)(), 1, j)
)
)
)
)(initial,array,Sumλ)tboulden I am going to have to work hard to catch up with you!! (grin)
Interesting discussion about thunking, have to learn more.
So far modified a bit my initial variant by adding inits to SCAN
scanByRows = LAMBDA(starts, data, fn,
MAKEARRAY(
ROWS(data),
COLUMNS(data),
LAMBDA(r, c,
// SCAN accumulates data for each row
// and outer INDEX returns result for
// each column of such row
INDEX(
SCAN(
INDEX(starts, r),
INDEX(data, r, 0),
LAMBDA(a, v, fn(a, v))
),
1,
c
)
)
)
);