Add head(), tail() and thin() methods?

[shoyer]

These would be shortcuts for isel/slice syntax:

• ds.head(time=5) -> ds.isel(time=slice(5)): select the first five time values
• ds.tail(time=5) -> ds.isel(time=slice(-5, None)): select the last five time values
• ds.thin(time=5) -> ds.isel(time=slice(None, None, 5)): select every 5th time value

[ebrevdo]

Clojure conventions: .take, .take_last get the first n and last n
pandas/ndarray conventions: .take([3,4,5]) selects rows 3,4,5.

probably want to be consistent with one of these.

On Tue, Feb 10, 2015 at 3:28 PM, Stephan Hoyer notifications@github.com
wrote:

These would be shortcuts for isel/slice syntax:

• ds.head(time=5) -> ds.isel(time=slice(5)): select the first five
  time values
• ds.tail(time=5) -> ds.isel(time=slice(-5, None)): select the last
  five time values
• ds.subsample(time=5) -> ds.isel(time=slice(None, None, 5)): select
  every 5th time value

—
Reply to this email directly or view it on GitHub
#319.

[shoyer]

@ebrevdo yes, I think we probably need to stick to the pandas/numpy convention for the meaning of take.

My inspiration for these names was the head() and tail() methods in pandas, which are quite convenient. But it's not entirely clear how/if these generalize cleanly to N-dimensions. I suppose take_first and take_last could be an improvement over head/tail.

[ebrevdo]

seems like for, e.g., head, you can pass either a single dimension or
multiple ones (e.g., either as **kwargs or a dictionary) and use those as
the start dimension.

that said, about naming conventions, i think for tensors the most common
convention is definitely slice() (which is implemented as isel). head/tail
can be implemented in terms of slice().

e.g.:
ds.slice(dim1=3, dim2=(1,4), dim3=(1,None,5)) -- or --
ds.slice({'dim1': 3, 'dim2': (1,4), 'dim3': (1, None, 5)})

head/tail/whatever are easy calls to this and you can have that in the
documentation. as a result, people won't get confused because they
understand slice.

On Tue, Feb 10, 2015 at 4:14 PM, Stephan Hoyer notifications@github.com
wrote:

@ebrevdo https://github.com/ebrevdo yes, I think we probably need to
stick to the pandas/numpy convention for the meaning of take.

My inspiration for these names was the head() and tail() methods in
pandas, which are quite convenient. But it's not entirely clear how/if
these generalize cleanly to N-dimensions. I suppose take_first and
take_last could be an improvement over head/tail.

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

[stale]

In order to maintain a list of currently relevant issues, we mark issues as stale after a period of inactivity

If this issue remains relevant, please comment here or remove the stale label; otherwise it will be marked as closed automatically

[jhamman]

I think this is still worth doing. I actually prefer @shoyer's original method names and API. I also think this would be a good first issue.