Expand description
The home screen’s ROWS preview: the first rows of the selected file, read the way
the open reads them and handed to the open as its first page (#547 M4).
File reads are not spent twice. The worker runs the open’s own scan and schema read and fills the page the table would ask for first; the pane shows a few rows of it, and Enter installs the dataset it built. The open then has nothing left to read for that page, and no scan or schema read of its own.
Structs§
- Prepared
- A dataset built and its first page read, waiting for the open that will show it.
- Preview
Rows - The first rows of a file, as text, for the pane: the leading columns and how many there are in all.
- Previews
- The previews read this session, and the one dataset waiting to be opened.
- Read
Counts - The reads of data the app has started, by kind: what a test counts to show a read was not made twice.
- Stamp
- What a file was when it was read: its size and modification time. A file that changed since is read again, and its old page is never installed.
Constants§
- PREVIEW_
ROWS - Rows the pane shows.
Functions§
- previewable
- Whether a row’s first rows are worth reading before it is opened: a local file of a
format whose first page is cheap, under
max_bytes(Parquet: its average row group). Nothing on a network share or in a store is read before it is opened.
Type Aliases§
- Handoff
- A prepared dataset passed from the preview to the open that takes it, once.