Usage with binary data ? about gota HOT 4 CLOSED

Hello, right now I'm working on storing arbitrary types inside the columns (A column can only have elements of the same type). This would allow to store lists or other data structures as cells inside the DataFrame matrix as long as they comply with the necessary interfaces.

How exactly do you want to use images or binary data with gota?

If it is for matrix computations I recommend you to take a look to https://github.com/gonum, for it is was conceived with that in mind.

Right now I'm focusing on data wrangling with this package, including things like subsetting, sorting, merging or function application but I don't discard adding more complex operations in the future.

from gota.

The reason I asked is because I need to stream binary data through
different computation nodes.
Use cases are real time video transformations, real time convolution &
analysis.

Gonum and its packages I presume can do the computation, but dataframes and
other buffered data structures are still needed to move the data between
the compute nodes.

Curious what you think.

On Sat, 13 Feb 2016, 14:47 Alex Sánchez [email protected] wrote:

Hello, right now I'm working on storing arbitrary types inside the columns
(A column can only have elements of the same type). This would allow to
store lists or other data structures as cells inside the DataFrame matrix
as long as they comply with the necessary interfaces.

How exactly do you want to use images or binary data with gota?

If it is for matrix computations I recommend you to take a look to
https://github.com/gonum, for it is was conceived with that in mind.

Right now I'm focusing on data wrangling with this package, including
things like subsetting, sorting, merging or function application but I
don't discard adding more complex operations in the future.

—
Reply to this email directly or view it on GitHub
#1 (comment).

from gota.

I suppose that you could implement your own type to store in the DataFrame cells, but keeping in mind your use cases I would say it makes more sense to use a less convoluted data structure: slice arrays, multidimensional arrays or hash maps should be able to do the job as long as you implement the real time algorithms correctly.

from gota.

Your right.. Will try out different way

On Sun, 14 Feb 2016, 01:39 Alex Sánchez [email protected] wrote:

I suppose that you could implement your own type to store in the DataFrame
cells, but keeping in mind your use cases I would say it makes more sense
to use a less convoluted data structure: slice arrays, multidimensional
arrays or hash maps should be able to do the job as long as you implement
the real time algorithms correctly.

—
Reply to this email directly or view it on GitHub
#1 (comment).

from gota.