Hi everyone

I am trying to create a matrix inside a matrix using a for loop.
This is what I did:

Matrix inside a matrix.sm (12 КиБ) скачан 31 раз(а).


This works but my real data is much larger than the one in this example and it takes like 10 minutes to run. Is there a more efficient way to do this?

Thanks.

Wrote

This works but my real data is much larger than the one in this example and it takes like 10 minutes to run. Is there a more efficient way to do this?

Your example runs in 0.5 s [Steam Engine Win 7]

Matrix inside a matrix.sm (12 КиБ) скачан 32 раз(а).

Two things. One move the assignment of p[k to under the k for loop. You're reassigning to p too often. Also, for large datasets, it helps to initialize the size of the matrix rather than just increasing the size over each iteration. See attaached. Matrix inside a matrix_lyles.sm (15 КиБ) скачан 32 раз(а).

Hola Pedro. Please check this version and tell us how long it takes with your data.

Matrix inside a matrix.sm (31 КиБ) скачан 36 раз(а).

Best regards.
Alvaro.

Wrote

Two things.

On my Win 7 ... takes 4 times my original SS 6179 doctored version.
Has it dependence with Smath version ?
Cheers ... Jean.

Wrote

Hola Pedro. Please check this version and tell us how long it takes with your data.

Timing is about twice my doctored version.

Wrote

Two things. One move the assignment of p[k to under the k for loop. You're reassigning to p too often. Also, for large datasets, it helps to initialize the size of the matrix rather than just increasing the size over each iteration. See attaached. Matrix inside a matrix_lyles.sm (15 КиБ) скачан 32 раз(а).

Thanks a lot, this works perfectly

Wrote

Hola Pedro. Please check this version and tell us how long it takes with your data.

Matrix inside a matrix.sm (31 КиБ) скачан 36 раз(а).

Best regards.
Alvaro.

Gracias, aprecio tu ayuda

Wrote

Your example runs in 0.5 s [Steam Engine Win 7]

... annotated attachment ...Cheers ... Jean

Matrix inside a matrix (Proposal).sm (31 КиБ) скачан 40 раз(а).

Wrote

This works but my real data is much larger than the one in this example and it takes like 10 minutes to run. Is there a more efficient way to do this?

1. Wait and see your next Monster Big.
2. 10 min on your machine might take 1 hr on mine, but willing.
3. Maybe not not so big and interpolate ?
See you sometimes, Cheers ... Jean