Hacktoberfest 2026:維護者為十月標記出來的 issue,仍然開放、適合新手。 瀏覽 Hacktoberfest issue

Reshaping Array

未關閉
#47 19 則留言 0 個 reaction 已指派 1 人 在 GitHub 檢視

@hpages 已經在處理了。

開始於 2019年7月9日。

評估

這個 Issue 還沒有評估資料。

描述

First, thanks for the package. It's exactly what we needed. We're working on neuroimaging data: https://github.com/muschellij2/niftiArray. We're trying to do statistics over multiple 3D arrays. I believe this doesn't fit well with DelayedMatrixStats as matrixStats requires 2D matrices.

Similar to https://github.com/Bioconductor/DelayedArray/issues/30 and the comment here https://github.com/Bioconductor/DelayedArray/issues/30#issuecomment-423683637, I was wondering about "reshaping" an array. Really, I want to be able to vectorize a DelayedArray or convert a 4D DelayedArray to a 2D matrix where the first 3 dimensions are the rows of the matrix and the 4th dimension is the columns. I can realize the array, convert to a matrix, and then create a DelayedArray or DelayedMatrix for the current solution.

Ideally, I'd be able to do this without realizing the array into memory. I'm not sure if HDF5 (the backend we're using) would support this, but we would it useful. Trying to set dim or change the dimensions fails due the (correct) checks on DelayedMatrix. Just wondering if it were possible, if it makes sense with HDF5, and if both are true, then would it be possible to implement? I can send a reprex if necessary, but the setting of dimensions is well documented.

主要語言
R
星號
29
分支
12
PR 合併指標
30 天內沒有已合併 PR

貢獻指南

這個儲存庫沒有索引到貢獻指南

從這裡開始

  1. 先讀完整個 Issue,再讀專案的貢獻指南。
  2. 在 Issue 下留言說明你要接手 —— 這能避免兩個人做同樣的事。
  3. Fork 儲存庫,在一個分支上完成修改。
  4. 送出 Pull Request,並在描述裡引用這個 Issue 編號。

Bioconductor/DelayedArray 的其他 Issue

查看 Bioconductor/DelayedArray 的全部 Issue

相似的 Issue

更多 R Issue

把新 issue 寄到你的電子郵件信箱

精選適合新手參與的 GitHub issue 摘要。