Re: [GENERAL] arrays of floating point numbers / linear algebra operations into the DB

[Enrico Sirola]

I respond myself:

Enrico Sirola ha scritto:

[...]

seems to work). The problem for the code above is that it doesn't work for vectors longer than 1000 elements or so (try it with 2000 and it doesn't work). I guess I should manage the "toasting" machinery in some ways - any suggestion is appreciated

wrong. it was just that I forgot to add ARR_OVERHEAD_NONULLS(ndims1) to
the mem allocation for rv:

rv = (ArrayType *) palloc(nbytes);

becomes

rv = (ArrayType *) palloc(nbytes) +  ARR_OVERHEAD_NONULLS(ndims1);

and now it seems to work :)

---------------------------(end of broadcast)---------------------------
TIP 6: explain analyze is your friend