Struct DatasetFacts 

pub struct DatasetFacts {
    pub mtime: u64,
    pub size: u64,
    pub rows: Option<usize>,
    pub cols: Option<usize>,
    pub cols_sampled: bool,
    pub columns: Vec<String>,
    pub kind: Option<EntryKind>,
    pub classified_by: u32,
    pub cost: Cost,
    pub holds: Holds,
}

What datui remembers about a dataset it has already measured.

This is a cache, not a catalogue. Every field is re-derivable by reading the dataset again, and each entry carries the size and modification time it was taken from, so a changed dataset invalidates itself. Deleting the file costs speed and nothing else — there is nothing here a user curated, and nothing that cannot be rebuilt by looking again.

Fields

mtime: u64

Modification time in seconds since the epoch, as a fingerprint.

size: u64

Size in bytes, the other half of the fingerprint.

rows: Option<usize>

cols: Option<usize>

cols_sampled: bool

Whether cols is a floor rather than a total: the directory was too large to read every footer of, so it was sampled. Restored with the count, or the row would present a sample as a total the next time it is listed.

columns: Vec<String>

Column names, which is what makes searching by column possible before anything has been read this run.

kind: Option<EntryKind>

What the dataset turned out to be. Recorded rather than re-derived, because a remote path cannot be classified without reading it, and a row that says hive under its own directory should not say something else under Recent.

classified_by: u32

Which build’s rules kind came from. See crate::discover::CLASSIFIER_VERSION. Absent in records written before this existed, which is what 0 means.

cost: Cost

What opening it will cost: compression, layout, partitioning. Worth keeping for the same reason the row count is — it came from a footer read that a remote dataset may not get a second chance at.

holds: Holds

What one listing of the directory found in it, which is what its label says. Restored beside kind and gated by the same classifier version: both are what looking into the directory produced, and a build that classified differently counted differently too.

Trait Implementations

impl Clone for DatasetFacts

fn clone(&self) -> Self

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for DatasetFacts

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for DatasetFacts

fn default() -> Self

Returns the “default value” for a type.

impl<'de> Deserialize<'de> for DatasetFacts

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Serialize for DatasetFacts

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.