Struct tidb_query_executors::table_scan_executor::TableScanExecutorImpl

[+] Show declaration

struct TableScanExecutorImpl {
    context: EvalContext,
    schema: Vec<FieldType>,
    columns_default_value: Vec<Vec<u8>>,
    column_id_index: HashMap<i64, usize>,
    handle_indices: SmallVec<[usize; 2]>,
    primary_column_ids: Vec<i64>,
    is_column_filled: Vec<bool>,
}

Fields

context: EvalContext

[−] Expand description

Note: Although called EvalContext, it is some kind of execution context instead.

schema: Vec<FieldType>

[−] Expand description

The schema of the output. All of the output come from specific columns in the underlying storage.

columns_default_value: Vec<Vec<u8>>

[−] Expand description

The default value of corresponding columns in the schema. When column data is missing, the default value will be used to fill the output.

column_id_index: HashMap<i64, usize>

[−] Expand description

The output position in the schema giving the column id.

handle_indices: SmallVec<[usize; 2]>

[−] Expand description

Vec of indices in output row to put the handle. The indices must be sorted in the vec.

primary_column_ids: Vec<i64>

[−] Expand description

Vec of Primary key column’s IDs.

is_column_filled: Vec<bool>

[−] Expand description

A vector of flags indicating whether corresponding column is filled in next_batch. It is a struct level field in order to prevent repeated memory allocations since its length is fixed for each next_batch call.

Implementations

impl TableScanExecutorImpl[−]

fn process_v1(
    &mut self,
    key: &[u8],
    value: &[u8],
    columns: &mut LazyBatchColumnVec,
    decoded_columns: &mut usize
) -> Result<()>

fn process_v2(
    &mut self,
    value: &[u8],
    columns: &mut LazyBatchColumnVec,
    decoded_columns: &mut usize
) -> Result<()>

Trait Implementations

impl ScanExecutorImpl for TableScanExecutorImpl[+]

[+] Show hidden undocumented items

fn schema(&self) -> &[FieldType][−]

Gets the schema.

fn mut_context(&mut self) -> &mut EvalContext[−]

Gets a mutable reference of the executor context.

fn build_column_vec(&self, scan_rows: usize) -> LazyBatchColumnVec[−]

Constructs empty columns, with PK in decoded format and the rest in raw format.

fn process_kv_pair(
    &mut self,
    key: &[u8],
    value: &[u8],
    columns: &mut LazyBatchColumnVec
) -> Result<()>[−]

Accepts a key value pair and fills the column vector.