Implement collect for int values

[datapythonista]

Not sure if this is related to #169.

Maybe I'm missing something, but I guess it'd be nice if the next code worked:

vec![1, 2, 3].iter().collect::<Int32Array>();

By the error I'm getting, I assume it's just not implemented. If you can confirm that this would be nice to add, I'm happy to give it a try (any hints welcome, I'm not very experienced with rust). This is the error:

$ cargo run 
   Compiling arrow-examples v0.1.0 (/home/mgarcia/src/arrow-examples)
error[E0277]: the trait bound `arrow::array::array_primitive::NativeAdapter<Int32Type>: From<&{integer}>` is not satisfied
    --> src/main.rs:18:26
     |
18   |     vec![1, 2, 3].iter().collect::<Int32Array>()
     |                          ^^^^^^^ the trait `From<&{integer}>` is not implemented for `arrow::array::array_primitive::NativeAdapter<Int32Type>`
     |
     = help: the following other types implement trait `From<T>`:
               <arrow::array::array_primitive::NativeAdapter<Float16Type> as From<half::binary16::f16>>
               <arrow::array::array_primitive::NativeAdapter<Float32Type> as From<f32>>
               <arrow::array::array_primitive::NativeAdapter<Float64Type> as From<f64>>
               <arrow::array::array_primitive::NativeAdapter<Int16Type> as From<i16>>
               <arrow::array::array_primitive::NativeAdapter<Int32Type> as From<i32>>
               <arrow::array::array_primitive::NativeAdapter<Int64Type> as From<i64>>
               <arrow::array::array_primitive::NativeAdapter<Int8Type> as From<i8>>
               <arrow::array::array_primitive::NativeAdapter<T> as From<&Option<<T as ArrowPrimitiveType>::Native>>>
             and 5 others
     = note: required because of the requirements on the impl of `Into<arrow::array::array_primitive::NativeAdapter<Int32Type>>` for `&{integer}`
     = note: required because of the requirements on the impl of `FromIterator<&{integer}>` for `PrimitiveArray<Int32Type>`
note: required by a bound in `collect`
    --> /home/mgarcia/.rustup/toolchains/stable-x86_64-unknown-linux-gnu/lib/rustlib/src/rust/library/core/src/iter/traits/iterator.rs:1788:19
     |
1788 |     fn collect<B: FromIterator<Self::Item>>(self) -> B
     |                   ^^^^^^^^^^^^^^^^^^^^^^^^ required by this bound in `collect`

For more information about this error, try `rustc --explain E0277`.
error: could not compile `arrow-examples` due to previous error

[tustvold]

The following will work

vec![1, 2, 3].into_iter().collect::<Int32Array>();
vec![1, 2, 3].iter().copied().collect::<Int32Array>();

I think this could be extended to support &i32, etc... by defining From<&i32> for NativeAdapter<i32>. Is this what you're after, or are the above solutions sufficient?

[datapythonista]

Thanks a lot for the fast answer @tustvold . I think your first example is what I want. I'm not so familiar with iterators, and similar examples that I saw used iter, not into_iter. Let me research a bit to understand better, but it should be enough.

[datapythonista]

I had a look, and the code you shared works great for integer, but for a string array seems it fails:

vec!["foo", "bar", "foobar"].iter().collect::<StringArray>()

   Compiling arrow-cookbook v0.1.0 (/home/mgarcia/src/arrow-cookbook)
error[E0277]: a value of type `GenericStringArray<i32>` cannot be built from an iterator over elements of type `&str`
    --> src/create_arrays.rs:90:46
     |
90   |     vec!["foo", "bar", "foobar"].into_iter().collect::<StringArray>()
     |                                              ^^^^^^^ value of type `GenericStringArray<i32>` cannot be built from `std::iter::Iterator<Item=&str>`
     |
     = help: the trait `FromIterator<&str>` is not implemented for `GenericStringArray<i32>`
     = help: the following other types implement trait `FromIterator<A>`:
               <GenericStringArray<OffsetSize> as FromIterator<&'a Option<Ptr>>>
               <GenericStringArray<OffsetSize> as FromIterator<Option<Ptr>>>
note: required by a bound in `collect`
    --> /home/mgarcia/.rustup/toolchains/stable-x86_64-unknown-linux-gnu/lib/rustlib/src/rust/library/core/src/iter/traits/iterator.rs:1788:19
     |
1788 |     fn collect<B: FromIterator<Self::Item>>(self) -> B
     |                   ^^^^^^^^^^^^^^^^^^^^^^^^ required by this bound in `collect`

For more information about this error, try `rustc --explain E0277`.
error: could not compile `arrow-cookbook` due to previous error

shell returned 101

Is this expected? Is there another syntax for this, or it just doesn't work for StringArray?

Thank you!

[tustvold]

We've not implemented a NativeAdapter for StringArray yet, so you can only collect an iterator of Option<&str>

So

["A',"b"].into_iter().map(Some).collect::<StringArray>()

[Some("A'),Some("b")].into_iter().map(Some).collect::<StringArray>()

Alternatively you can use https://docs.rs/arrow/22.0.0/arrow/array/struct.GenericStringArray.html#method.from_iter_values

[datapythonista]

Ah, I see. Thanks a lot for the help. I may give it a try and see if I can implement it. Closing this issue.