Add nodata example to collections

[manycoding]

For me it seems the most useful example. In particular, if we want to parallelize collections download we can first fetch all keys then split on batches.
I would also like to add nodata to both items and collectionsiter(), but I don't know how to do it.

[manycoding]

Closes #124

[Gallaecio]

For this, wouldn’t it be slightly better to use .iter() with https://docs.python.org/3/library/itertools.html#itertools.islice, to avoid building a big list?

[vshlapakov]

If your list has only 1k elements (and it's important because iter & list handlers return up to 1k elements max by default), this won't work:

>>> next(itertools.islice([1,2,3], 3, 4))
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
StopIteration

Also, getting Nth item's key is simpler (and more lightweight) by providing index param or getting item by the index, isn't it? In this case, you don't need to go through elements at all. There might be additional filters, so it probably won't work as you want in all cases, but as a general approach for an abstract scenario (i.e. as an example in the docs) that's a way better.

>>> job.items.list(index=1000)
{'field': 'data', ...}

>>> job.items.get(1000)
{'field': 'data', ...}

I'd suggest you to pick another example to demonstrate the nodata feature, smth like:

- iterate over keys only (doesn't fetch items data)

    >>> for elem in foo_store.iter(nodata=True, meta=["_key"])):
    ...     print(elem)
    {'_key': '002d050ee3ff6192dcbecc4e4b4457d7'}

[manycoding]

I thought it's obvious that we know our dataset size is bigger then 1000, isn't it?

(and it's important because iter & list handlers return up to 1k elements max by default)

It seems they return all elements by default, and islice is the fastest way https://jupyter.scrapinghub.com/user/v/lab/tree/shared/Experiments/Arche/collections_nodata.ipynb

But I made the example simpler.

[vshlapakov]

I thought it's obvious that we know our dataset size is bigger then 1000, isn't it?

Maybe, but it's very common that users are glancing through the examples very fluently, and their assumptions can be totally different 🙂 Thanks!

[vshlapakov]

Thanks @manycoding 👍