Struct TableCatalog

pub struct TableCatalog {

    pub id: TableId,
    pub associated_source_id: Option<TableId>,
    pub name: String,
    pub dependent_relations: Vec<TableId>,
    pub columns: Vec<ColumnCatalog>,
    pub pk: Vec<ColumnOrder>,
    pub stream_key: Vec<usize>,
    pub table_type: TableType,
    pub distribution_key: Vec<usize>,
    pub append_only: bool,
    pub cardinality: Cardinality,
    pub owner: u32,
    pub retention_seconds: Option<u32>,
    pub fragment_id: u32,
    pub dml_fragment_id: Option<u32>,
    pub vnode_col_index: Option<usize>,
    pub row_id_index: Option<usize>,
    pub value_indices: Vec<usize>,
    pub definition: String,
    pub conflict_behavior: ConflictBehavior,
    pub version_column_index: Option<usize>,
    pub read_prefix_len_hint: usize,
    pub version: Option<TableVersion>,
    pub watermark_columns: FixedBitSet,
    pub dist_key_in_pk: Vec<usize>,
    pub created_at_epoch: Option<Epoch>,
    pub initialized_at_epoch: Option<Epoch>,
    pub cleaned_by_watermark: bool,
    pub create_type: CreateType,
    pub stream_job_status: StreamJobStatus,
    pub description: Option<String>,
    pub incoming_sinks: Vec<u32>,
    pub created_at_cluster_version: Option<String>,
    pub initialized_at_cluster_version: Option<String>,
    pub cdc_table_id: Option<String>,
    pub vnode_count: VnodeCount,
}

TableCatalog Includes full information about a table.

Here Table is an internal concept, corresponding to a table in storage, all of which can be SELECTed. It is not the same as a user-side table created by CREATE TABLE.

Use Self::table_type() to determine the TableType of the table.

Column ID & Column Index

ColumnId (with type i32) is the unique identifier of a column in a table. It is used to access storage.

Column index, or idx, (with type usize) is the relative position inside the Vec of columns.

A tip to avoid making mistakes is never do casting - i32 as usize or vice versa.

Keys

All the keys are represented as column indices.

• Primary Key (pk): unique identifier of a row.

• Order Key: the primary key for storage, used to sort and access data.

  For an MV, the columns in ORDER BY clause will be put at the beginning of the order key. And the remaining columns in pk will follow behind.

  If there’s no ORDER BY clause, the order key will be the same as pk.

• Distribution Key: the columns used to partition the data. It must be a subset of the order key.

Fields

id: TableId

associated_source_id: Option<TableId>

name: String

dependent_relations: Vec<TableId>

columns: Vec<ColumnCatalog>

All columns in this table.

pk: Vec<ColumnOrder>

Key used as materialize’s storage key prefix, including MV order columns and stream_key.

stream_key: Vec<usize>

pk_indices of the corresponding materialize operator’s output.

table_type: TableType

Type of the table. Used to distinguish user-created tables, materialized views, index tables, and internal tables.

distribution_key: Vec<usize>

Distribution key column indices.

append_only: bool

The append-only attribute is derived from StreamMaterialize and StreamTableScan relies on this to derive an append-only stream plan.

cardinality: Cardinality

The cardinality of the table.

owner: u32

Owner of the table.

retention_seconds: Option<u32>

fragment_id: u32

The fragment id of the Materialize operator for this table.

dml_fragment_id: Option<u32>

The fragment id of the DML operator for this table.

vnode_col_index: Option<usize>

An optional column index which is the vnode of each row computed by the table’s consistent hash distribution.

row_id_index: Option<usize>

An optional column index of row id. If the primary key is specified by users, this will be None.

value_indices: Vec<usize>

The column indices which are stored in the state store’s value with row-encoding.

definition: String

The full CREATE TABLE or CREATE MATERIALIZED VIEW definition of the table.

conflict_behavior: ConflictBehavior

The behavior of handling incoming pk conflict from source executor, we can overwrite or ignore conflict pk. For normal materialize executor and other executors, this field will be No Check.

version_column_index: Option<usize>

read_prefix_len_hint: usize

version: Option<TableVersion>

Per-table catalog version, used by schema change. None for internal tables and tests.

watermark_columns: FixedBitSet

The column indices which could receive watermarks.

dist_key_in_pk: Vec<usize>

Optional field specifies the distribution key indices in pk. See https://github.com/risingwavelabs/risingwave/issues/8377 for more information.

created_at_epoch: Option<Epoch>

initialized_at_epoch: Option<Epoch>

cleaned_by_watermark: bool

Indicate whether to use watermark cache for state table.

create_type: CreateType

Indicate whether to create table in background or foreground.

stream_job_status: StreamJobStatus

Indicate the stream job status, whether it is created or creating. If it is creating, we should hide it.

description: Option<String>

description of table, set by comment on.

incoming_sinks: Vec<u32>

Incoming sinks, used for sink into table

created_at_cluster_version: Option<String>

initialized_at_cluster_version: Option<String>

cdc_table_id: Option<String>

vnode_count: VnodeCount

Total vnode count of the table.

Can be VnodeCount::Placeholder if the catalog is generated by the frontend and the corresponding job is still in Creating status, in which case calling Self::vnode_count will panic.

Implementations

impl TableCatalog

pub fn id(&self) -> TableId

Get a reference to the table catalog’s table id.

pub fn with_id(self, id: TableId) -> Self

pub fn with_cleaned_by_watermark(self, cleaned_by_watermark: bool) -> Self

pub fn conflict_behavior(&self) -> ConflictBehavior

pub fn table_type(&self) -> TableType

pub fn is_table(&self) -> bool

pub fn is_internal_table(&self) -> bool

pub fn is_mview(&self) -> bool

pub fn is_index(&self) -> bool

pub fn bad_drop_error(&self) -> RwError

Returns an error if DROP statements are used on the wrong type of table.

pub fn associated_source_id(&self) -> Option<TableId>

Get the table catalog’s associated source id.

pub fn has_associated_source(&self) -> bool

pub fn columns(&self) -> &[ColumnCatalog]

Get a reference to the table catalog’s columns.

pub fn columns_without_rw_timestamp(&self) -> Vec<ColumnCatalog>

pub fn pk(&self) -> &[ColumnOrder]

Get a reference to the table catalog’s pk desc.

pub fn pk_column_ids(&self) -> Vec<ColumnId>

Get the column IDs of the primary key.

pub fn table_desc(&self) -> TableDesc

Get a TableDesc of the table.

Note: this must be called on existing tables, otherwise it will fail to get the vnode count (which is determined by the meta service) and panic.

pub fn name(&self) -> &str

Get a reference to the table catalog’s name.

pub fn distribution_key(&self) -> &[usize]

pub fn to_internal_table_prost(&self) -> PbTable

pub fn create_sql(&self) -> String

Returns the SQL statement that can be used to create this table.

pub fn version(&self) -> Option<&TableVersion>

Get a reference to the table catalog’s version.

pub fn version_id(&self) -> Option<TableVersionId>

Get the table’s version id. Returns None if the table has no version field.

pub fn vnode_count(&self) -> usize

Get the total vnode count of the table.

Panics if it’s called on an incomplete (and not yet persisted) table catalog.

pub fn to_prost(&self, schema_id: u32, database_id: u32) -> PbTable

pub fn columns_to_insert(&self) -> impl Iterator<Item = &ColumnCatalog>

Get columns excluding hidden columns and generated golumns.

pub fn generated_column_names(&self) -> impl Iterator<Item = &str>

pub fn generated_col_idxes(&self) -> impl Iterator<Item = usize> + '_

pub fn default_column_expr(&self, col_idx: usize) -> ExprImpl

pub fn default_column_exprs(columns: &[ColumnCatalog]) -> Vec<ExprImpl>

pub fn default_columns(&self) -> impl Iterator<Item = (usize, ExprImpl)> + '_

pub fn has_generated_column(&self) -> bool

pub fn has_rw_timestamp_column(&self) -> bool

pub fn column_schema(&self) -> Schema

pub fn is_created(&self) -> bool

Trait Implementations

impl Clone for TableCatalog

fn clone(&self) -> TableCatalog

Returns a copy of the value.

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

Performs copy-assignment from source.

impl Debug for TableCatalog

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

Formats the value using the given formatter.

impl From<&Table> for TableCatalog

fn from(tb: &PbTable) -> Self

Converts to this type from the input type.

impl From<Table> for TableCatalog

fn from(tb: PbTable) -> Self

Converts to this type from the input type.

impl Hash for TableCatalog

fn hash<__H: Hasher>(&self, state: &mut __H)

Feeds this value into the given Hasher.

fn hash_slice<H>(data: &[Self], state: &mut H)

where H: Hasher, Self: Sized,

Feeds a slice of this type into the given Hasher.

impl OwnedByUserCatalog for TableCatalog

fn owner(&self) -> u32

Returns the owner of the catalog.

impl PartialEq for TableCatalog

fn eq(&self, other: &TableCatalog) -> bool

Tests for self and other values to be equal, and is used by ==.

fn ne(&self, other: &Rhs) -> bool

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl Eq for TableCatalog

impl StructuralPartialEq for TableCatalog