NaN as a result

[YA_10963798]

I don't know why I getting this kind of result.. I tried everything. It is on page 15 . I use prime 9

Thanks in advance for your help

[StuartBruff]

---

@YA_10963798 wrote:

I don't know why I getting this kind of result.. I tried everything. It is on page 15 . I use prime 9

---

When you multiply any pair of vectors that contain at least one NaN, the result is also a NaN.

Stuart

I've got Mathcad Prime 10, so I can't post a worksheet that's useful to you.  Sorry.

[YA_10963798]

so that means the whole results will be NaN.. and that will erase all other results?.

[StuartBruff]

Yes, I’m afraid so.

Stuart

[YA_10963798]

thank you for trying . I appreciate your time 

[Werner_E · Accepted solution]

Whenever in  any calculation a NaN is involved, the result is NaN.

Maybe a utility  function to do the vector multiplication ignoring any multiplications with NaN involved could help:

Applied to your calculation it looks like this:

Is this somewhere near to what you expected?

The calculation takes some time as Prime has to multiply 3669*2=7338 times two 3670x1 vectors  and a user-written function using augment and filterNaN isn't very fast.

Prime 9 sheet attached

[YA_10963798]

Thank you that solved my problem

[StuartBruff]

---

@Werner_E wrote:

Whenever in  any calculation a NaN is involved, the result is NaN.

Maybe a utility  function to do the vector multiplication ignoring any multiplications with NaN involved could help:

The calculation takes some time as Prime has to multiply 3669*2=7338 times two 3670x1 vectors  and a user-written function using augment and filterNaN isn't very fast.

---

Out of idle curiosity, I wondered what you meant by "takes some time".  Now I know.

Surprisingly, a brute force and ignorance method offers better performance.

Stuart

[Werner_E]

Not such a big surprise. I suspect that "filterNaN" is the main bottleneck and not "augment" or the column selector.

The if function seems to be faster even though its used in the loop.

[StuartBruff]

---

@Werner_E wrote:

Not such a big surprise. I suspect that "filterNaN" is the main bottleneck and not "augment" or the column selector.

The if function seems to be faster even though its used in the loop.

---

Indeed, filterNaN does seem to be the larger bottleneck.  However, augment is not entirely innocent in these proceedings.

Stuart