René's URL Explorer Experiment

{"@context":"https://schema.org","@type":"DiscussionForumPosting","headline":"Zip and azip! behavior related to inputted NdProducers","articleBody":"Based on trying to help in Issue #452 , I found some interesting behavior regarding either `Zip` or `azip!`. I've given examples down below\r\n\r\n```\r\n    let a = arr3(\u0026[[[ 0,  1,  2],\r\n                [ 3,  4,  5]],\r\n               [[ 6,  7,  8],\r\n                [ 9, 10, 11]]]);\r\n    let inner0 = a.lanes(Axis(0));\r\n\r\n    let trial = arr1(\u0026[0,0,0,0,0,0]);\r\n    let tr_iter = trial.axis_iter(Axis(0));\r\n\r\n    //The below will error out when compiling.\r\n    azip!(ref output (inner0), ref trial (tr_iter) in {println!(\"{:?}, {:?}\",output, trial); });\r\n```\r\nThe compiler returns the following error:\r\n\r\n```\r\nerror[E0271]: type mismatch resolving `\u003cndarray::iter::AxisIter\u003c'_, {integer}, ndarray::Dim\u003c[usize; 0]\u003e\u003e as ndarray::IntoNdProducer\u003e::Dim == ndarray::Dim\u003c[usize; 2]\u003e`\r\n   --\u003e src/main.rs:141:5\r\n    |\r\n141 |     azip!(ref output (inner0), ref trial (tr_iter) in {println!(\"{:?}, {:?}\",output, trial); });\r\n    |     ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ expected an array with a fixed size of 1 elements, found one with 2 elements\r\n    |\r\n    = note: expected type `ndarray::Dim\u003c[usize; 1]\u003e`\r\n               found type `ndarray::Dim\u003c[usize; 2]\u003e`\r\n    = note: this error originates in a macro outside of the current crate\r\n\r\nerror[E0599]: no method named `apply` found for type `ndarray::Zip\u003c(ndarray::iter::Lanes\u003c'_, {integer}, ndarray::Dim\u003c[usize; 2]\u003e\u003e, ndarray::iter::AxisIter\u003c'_, {integer}, ndarray::Dim\u003c[usize; 0]\u003e\u003e), ndarray::Dim\u003c[usize; 2]\u003e\u003e` in the current scope\r\n   --\u003e src/main.rs:141:5\r\n    |\r\n141 |     azip!(ref output (inner0), ref trial (tr_iter) in {println!(\"{:?}, {:?}\",output, trial); });\r\n    |     ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\r\n    |\r\n    = note: this error originates in a macro outside of the current crate\r\n```\r\n\r\n\r\nHowever, if I were to roll what inner0 outputs directly, I can get it to work no problem as seen below:\r\n\r\n\r\n```\r\n    let trial = arr1(\u0026[0,0,0,0,0,0]);\r\n    let tr_iter = trial.axis_iter(Axis(0));\r\n\r\n    let b = arr2(\u0026[[0,1,2,3,4,5],[6,7,8,9,10,11]]);\r\n    let b_iter = b.axis_iter(Axis(1));\r\n    //The compiler is perfectly okay with this code.\r\n    azip!(ref output (b_iter), ref trial (tr_iter) in {println!(\"{:?}, {:?}\",output, trial); });\r\n```\r\n\r\nI feel that either method should work, since we contain the same number of \"items\" to iterate over in each version. So, I was curious if this is simply an error/bug in the code or is there a design choice behind one method failing to compile and the other one working.","author":{"url":"https://github.com/rcarson3","@type":"Person","name":"rcarson3"},"datePublished":"2018-05-17T22:32:11.000Z","interactionStatistic":{"@type":"InteractionCounter","interactionType":"https://schema.org/CommentAction","userInteractionCount":7},"url":"https://github.com/453/ndarray/issues/453"}