Kinetica Virtual Catalog

Overview

The Kinetica Virtual Catalog tables contain metadata about the database's objects, relationships, & permissions.

The tables are implemented commensurate with their usage and storage requirements. As such, a variety of distribution and partitioning schemes will be used across the tables. These tables will be automatically updated to reflect changes in the database structure and processing state.

The information returned when querying the virtual catalogs is automatically filtered by the permissions of the querying user. For instance, regular users will only see the tables & views to which they have access when querying a virtual catalog table that lists all of the tables in the database, while system administrators will see all tables & views.

For queries that can be run against the catalog tables to return useful information about the structure and processing state of the database, see Useful Catalog Queries.

The Kinetica Virtual Catalog tables reside in the ki_catalog schema.

Catalog Table List

Virtual Catalog Table	Description
ki_columns	Columns of tables & views and their respective types & statistics.
ki_contexts	SQL-GPT context entries and their configuration. For each context, there will be one record per section--one for each table entry, one for global rules, and one for global samples.
ki_datasource_subscriptions	Kafka data subscriptions, associating Kafka data sources with their target load tables.
ki_datasources	External data sources, their containing schemas and types.
ki_datatypes	Supported column data types.
ki_depend	Object dependencies for views, procedures, and result tables.
ki_fk_constraints	Foreign keys, one record per column specified in each foreign key.
ki_functions	Scalar, aggregate, window, and user-defined functions.
ki_ingest_file_info	Name & date of data files loaded into the database.
ki_load_history	Record & statistics of each data load or export, subscription event, and data refresh.
ki_obj_stat	Row & byte counts for each object in the database.
ki_object_permissions	Privileges granted, including the grantee, privilege type, and object to which access was granted.
ki_objects	Tables & views and their respective types & statistics.
ki_query_active_all	Actively running SQL statements per worker rank, including DML & DDL.
ki_query_history	History of SQL statements run within the database, including DML & DDL; will not include multi-head ingest/egress.
ki_query_span_metrics_all	Processing metrics captured during the execution of commands on worker ranks.
ki_query_workers	Worker ranks' status and their actively running tasks.
ki_rag_embeddings	SQL-GPT context definitions and their associated embeddings, one record for each sample or table definition.
ki_role_members	Users & roles and their granted roles, one record per user/role or role/role pair.
ki_schemas	Schemas, containing all database objects.
ki_tiered_objects	Tiered objects and their size within those tiers. Tip To retrieve the tiering information for a particular object, the object must be queried by fully-qualified name using the pseudo-column `outer_object`.
ki_users_and_roles	Users & user roles.

Catalog Column List

ki_columns

Each record in ki_columns represents a single column in a table or view for all tables & views in the database, including those in virtual catalog tables.

Column Name	Column Type	Description
`oid`	`BIGINT`	Unique identifier associated with this column, database-wide.
`schema_oid`	`BIGINT`	Unique identifier associated with the schema containing this column's table. See ki_schemas.
`schema_name`	`VARCHAR(256)`	Name of the schema containing this column's table.
`table_oid`	`BIGINT`	Unique identifier associated with the table containing this column. See ki_objects.
`table_name`	`VARCHAR(256)`	Name of the table containing this column.
`column_name`	`VARCHAR(256)`	Name of this column.
`column_position`	`INTEGER`	Position of this column in its table, using a 0-based index.
`column_type_oid`	`BIGINT`	Unique identifier associated with the data type of this column. See ki_datatypes.
`is_nullable`	`INTEGER`	Whether this column can be set to null.
`is_shard_key`	`INTEGER`	Whether this column is a shard key or part of a composite shard key.
`is_primary_key`	`INTEGER`	Whether this column is a primary key or part of a composite primary key.
`is_dist_encoded`	`INTEGER`	Whether this column is dictionary-encoded.
`default_value`	`VARCHAR(256)`	Not used.
`default_value_expr`	`VARCHAR(256)`	The default value expression applied to this column. See Data Replacement for options.
`properties`	`VARCHAR(256)`	Space-separated list of all column properties applied to this column.
`comments`	`VARCHAR(256)`	User comments associated with this column.

ki_contexts

Each record in ki_contexts represents a top-level object within a SQL-GPT context. Objects can be one of the following:

Table Definition - Rules, samples & comments that apply to the use of the given table specifically
Rules Definition - Rules that apply globally when using the context
Samples Definition - Samples that apply globally when using the context

Column Name Column Type Description

oid BIGINT Unique identifier associated with this SQL-GPT context object.

context_name VARCHAR(256) Name of the SQL-GPT context containing this context object.

schema_oid BIGINT Unique identifier associated with the schema containing the parent SQL-GPT context. See ki_schemas.

schema_name VARCHAR(256) Name of the schema containing the parent SQL-GPT context.

object_name

VARCHAR(256)

Name of this SQL-GPT context object, denoting the type of context object it is.

Name	Description
`rules`	Rules Definition
`samples`	Samples Definition
<schema_name.table_name>	Table Definition for the given table name

object_description VARCHAR The COMMENT associated with the table, if context object is a table definition.

object_rules VARCHAR[] The RULES associated with the table, if context object is a table definition; or the global rules associated with the overall context, if a rules definition.

object_comments JSON The COMMENTS associated with the table, if context object is a table definition--essentially, the column definitions for the table.

object_samples JSON The SAMPLES associated with the table, if context object is a table definition; or the global samples associated with the overall context, if a samples definition.

is_temp_context BOOLEAN Whether this context is temporary and will be removed when either its time-to-live expires or the database is restarted, whichever comes first.

ttl INTEGER Time-to-live, in seconds, for this context; if this context is not used for the associated TTL time span, it will be automatically removed.

ki_datasource_subscriptions

Each record in ki_datasource_subscriptions represents a subscription a single target table has through a single data source.

Column Name	Column Type	Description
`datasource_oid`	`BIGINT`	Unique identifier associated with the data source used by this subscription. See ki_datasources.
`table_oid`	`BIGINT`	Unique identifier associated with the table to which subscribed data is written. See ki_objects.

ki_datasources

Each record in ki_datasources represents a single data source.

Column Name	Column Type	Description
`datasource_oid`	`BIGINT`	Unique identifier associated with this data source.
`datasource_schema`	`VARCHAR(256)`	Name of the schema containing this data source.
`datasource_name`	`VARCHAR(256)`	Name of this data source.
`datasource_kind`	`VARCHAR(8)`	Provider associated with this data source.

ki_datatypes

Each record in ki_datatypes represents a standard data type into which a supported data type can be translated. Not all supported data types will have a unique 1-to-1 mapping with the standard set, and not all data types listed in this table will have supported analogs in Kinetica.

Column Name	Column Type	Description
`oid`	`BIGINT`	Unique identifier associated with this data type.
`name`	`VARCHAR(256)`	Name of the data type.
`schema_oid`	`BIGINT`	Unique identifier associated with the schema containing this data type. See ki_schemas.
`creator`	`VARCHAR(256)`	Not used.
`kind`	`VARCHAR(1)`	Category of this data type; always `B` for Base type.
`size`	`INTEGER`	Size, in bytes, of this data type; `-1` indicates variable length. Note that fixed-width (charN) string columns will list their actual data type size in ki_columns.
`is_varying`	`INTEGER`	Whether the data type is variable length.
`prec`	`INTEGER`	Not used.
`scale`	`INTEGER`	Not used.
`pg_typename`	`VARCHAR(32)`	Name of the equivalent PostgreSQL type.
`sql_typename`	`VARCHAR(32)`	Name of the equivalent ANSI SQL type.

ki_depend

Each record in ki_depend represents a dependent relationship between a "source" object and an object that depends on it; for example, a view will depend on all of the tables & views referenced in its corresponding query.

Column Name Column Type Description

src_obj_oid BIGINT Unique identifier associated with the source object depended on.

src_obj_kind

VARCHAR(1)

Type of the source object depended on.

Code	Description
E	External table
H	Partitioned table
I	Materialized view intermediate table
M	Materialized view
P	SQL procedure
R	Regular (local) table
V	Logical view

dep_obj_oid BIGINT Unique identifier associated with the dependent object.

dep_obj_kind

VARCHAR(1)

Type of the dependent object.

Code	Description
I	Materialized view intermediate table
M	Materialized view
P	SQL procedure
R	Regular (local) table
V	Logical view

mv_oid BIGINT Unique identifier of the associated materialized view (for materialized view components).

dep_kind VARCHAR(1) Not used.

ki_fk_constraints

Each record in ki_fk_constraints represents one column in either a primary key or foreign key relationship. For example, a table with a single primary key column will have one record in this table, while a composite foreign key linking two columns of one table to two columns of another will have four records in this table--one for each of the columns on the source side of the relation and one for each of the columns on the target side of the relation. Primary keys will be listed once for themselves, and then once for each instance of their participation in a foreign key relationship.

Column Name Column Type Description

schema_oid BIGINT Unique identifier associated with the schema containing this constrained table. See ki_schemas.

schema_name VARCHAR(256) Name of the schema containing this constrained table.

table_oid BIGINT Unique identifier associated with the table containing this constraint. See ki_objects.

table_name VARCHAR(256) Name of the table containing this constraint.

parent_schema_oid BIGINT Unique identifier associated with the schema containing the foreign table referenced by this constraint (for foreign key constraints). See ki_schemas.

parent_schema_name VARCHAR(256) Name of the schema containing the foreign table referenced by this constraint (for foreign key constraints).

parent_table_oid BIGINT Unique identifier associated with the foreign table referenced by this constraint (for foreign key constraints). See ki_objects.

parent_table_name VARCHAR(256) Name of the foreign table referenced by this constraint (for foreign key constraints).

constraint_name VARCHAR(256) Name of this constraint, if specified by the user (for foreign key constraints).

unique_constraint_name VARCHAR(256) Name of this constraint, auto-generated by the system.

created_by VARCHAR(256) Username of user who created the constraint.

creation_time TIMESTAMP Timestamp of constraint creation.

last_alter_time TIMESTAMP Timestamp of most recent constraint update (initially, time of creation).

is_enforced

INTEGER

Whether the constraint is enforced.

Code	Description
0	Ignored (all foreign key constraints are ignored)
1	Enforced (all primary key constraints are enforced)

fk_column_count INTEGER Number of columns in the associated key.

fk_column_name VARCHAR(256) Name of this constrained column referencing the primary key column of the foreign table.

fk_column_pos INTEGER 0-based index of the position of this constrained column within the set of foreign key columns (for foreign key constraints). If the foreign key is composed of only one column (non-composite foreign key), the index will be 0.

pk_column_name VARCHAR(256) Name of the primary key column in the foreign table referenced by this constrained column (for foreign key constraints).

pk_column_pos INTEGER 0-based index of the position of the primary key column in the foreign table referenced by this constrained column within the set of referenced primary key columns (for foreign key constraints). If the primary key in the foreign table is composed of only one column (non-composite primary key), the index will be 0.

fk_column_positions INTEGER[] 0-based indexes of all foreign key columns associated with this constraint within the set of all columns in the constrained table (for foreign key constraints).

pk_column_positions INTEGER[] 0-based indexes of all primary key columns associated with this constraint within the set of all columns in the constrained table. For primary key constraints, this will be the list of column indexes among all columns in the constrained table; for foreign key constraints, this will be the list of column indexes among all columns in the foreign table referenced by the constraint.

fk_def VARCHAR(256) Definition of the foreign key constraint (for foreign key constraints).

pk_def VARCHAR(256) Definition of the primary key constraint. For primary key constraints, this will be the definition of the primary key on the constrained table; for foreign key constraints, this will be the definition of the primary key on the foreign table referenced by the constraint.

comments VARCHAR(256) Not used.

ki_functions

Each record in ki_functions represents a native scalar or aggregate function, a UDF/UDTF, or a SQL procedure.

Column Name Column Type Description

oid BIGINT Unique identifier associated with this function.

schema_oid BIGINT Unique identifier associated with the schema containing this function. See ki_schemas.

schema_name VARCHAR(256) Name of the schema containing this constrained table. Native functions and UDFs are located in the pg_catalog schema, while SQL procedures are located in the schema specified by the user during creation.

name VARCHAR(256) Name of this function.

lang VARCHAR(256) Language in which this SQL procedure was written (for SQL procedures; always SQL).

variadic INTEGER Not used.

kind

VARCHAR(256)

Function type.

Code	Description
A	Aggregate function
C	Cast function
P	SQL procedure
S	Scalar function
U	User-defined function

volatile VARCHAR(256) Not used.

nargs INTEGER Number of parameters accepted by this function.

default_arg_count INTEGER Number of parameters accepted by this function with default values.

return_type_oid BIGINT Unique identifier associated with the data type of this function's return value. See ki_datatypes.

arg_type_oid BIGINT[] Unique identifiers associated with the data types of this function's parameters. See ki_datatypes.

arg_types VARCHAR[] Names of the data types of this function's parameters. See ki_datatypes.

arg_names VARCHAR Not used.

creation_time TIMESTAMP Timestamp of function creation.

last_alter_time TIMESTAMP Timestamp of most recent function update (initially, time of creation).

created_by VARCHAR(256) Username of user who created this function.

properties VARCHAR(256) Not used.

info VARCHAR(256) Not used.

ki_ingest_file_info

Each record in ki_ingest_file_info represents the import from a single data file into the database or the export into a single data file from the database.

Column Name	Column Type	Description
`file_oid`	`BIGINT`	Unique identifier associated with the data file imported from or exported into.
`jobid`	`BIGINT`	Unique identifier of the system job responsible for processing the import/export request.
`start_time`	`TIMESTAMP`	Timestamp of import/export start.
`file_short_name`	`VARCHAR(64)`	Name of data file.
`file_full_name`	`VARCHAR`	Full path of data file.

ki_load_history

Each record in ki_load_history represents a process that was invoked during either the import into or export from the database. Multiple processes may be invoked for a single data processing task.

Column Name Column Type Description

table_oid BIGINT Unique identifier associated with the table into which loaded data is inserted. See ki_objects.

datasource_oid BIGINT Unique identifier associated with the data source or data sink used for this load (for loads that make use of either). See ki_datasources for loads that use a data source.

user_name VARCHAR(256) Name of the user who initiated the load.

jobid BIGINT Unique identifier of the system job responsible for the load.

rank_num INTEGER Index of the rank responsible for processing this segment of the load.

tom INTEGER Index of the TOM responsible for processing this segment of the load.

load_kind

VARCHAR(1)

Type of load.

Code	Description
E	Export local table to remote table via data sink.
I	Ingest from KiFS file or remote data via data source.
X	Export local table to remote file via data sink.

start_time TIMESTAMP Timestamp of loading process start.

end_time TIMESTAMP Timestamp of loading process end.

file_oid BIGINT Unique identifier associated with the file imported from or exported into (for loads that make use of files). See ki_ingest_file_info.

line_num BIGINT Not used.

offset_val BIGINT Not used.

rows_processed BIGINT Number of records processed for this import/export.

rows_inserted BIGINT Number of records processed for this load that were either successfully imported into Kinetica or exported into a remote database (for imports of any kind and exports to tables).

rows_skipped BIGINT Number of records processed for this import that were skipped due to errors in the data (for imports).

event_message

VARCHAR

System messages relevant to the load process, including:

error message relating to skipped records or other, if applicable (for imports)
Cancelled, denoting a data load process was terminated
pause, denoting a data load subscription was stopped temporarily
resume, denoting a data load subscription was resumed after being paused
subscribe, denoting subscription-based data loads
unsubscribe, denoting a data load subscription was cancelled

num_errors BIGINT Number of errors encountered during processing.

additional_info

VARCHAR

Extra information about the load, depending on load type.

Load Type	Additional Info
Import from file	Indicator of which import load process this was.
Import from query	Remote query used to pull data from remote database.
Export to table	How the export query was split across the local table.

ki_obj_stat

Each record in ki_obj_stat represents a single table or view in the database and its associated sizing statistics, including those in virtual catalog tables.

Column Name	Column Type	Description
`oid`	`BIGINT`	Unique identifier associated with this object.
`object_name`	`VARCHAR(256)`	Name of this object.
`schema_name`	`VARCHAR(256)`	Name of the schema containing this object.
`row_count`	`BIGINT`	Number of records contained in this object.
`bytes_per_row`	`BIGINT`	Calculated average number of bytes per record for this object.
`total_bytes`	`BIGINT`	Total number of raw bytes used by the table; this will not account for the extra RAM occupied by the unused portion of data chunks containing this table or the blocks used on disk to store the table.

ki_object_permissions

Each record in ki_object_permissions represents the mapping of a user or role to a system-level or object-level permission they have.

Column Name Column Type Description

grantor BIGINT Not used.

role_oid BIGINT Unique identifier of the user or role associated with this mapping. See ki_users_and_roles.

role_name VARCHAR(256) Name of the user or role associated with this mapping.

permission_type VARCHAR(256) Type of permission grant this mapping represents.

object_type

VARCHAR(256)

Type of object grant this mapping represents (for object-level permissions).

Type	Description
`proc`	SQL procedure or UDF/UDTF
`datasink`	Data sink
`datasource`	Data source
`credential`	Credential
`directory`	KiFS directory
`schema`	Schema
`graph`	Graph
`context`	SQL-GPT context
`table`	Table or view
`table_monitor`	Table monitor

object_schema_oid BIGINT Unique identifier of the schema containing the granted object associated with this mapping. See ki_schemas.

object_schema_name VARCHAR(256) Name of the schema containing the granted object associated with this mapping.

object_oid BIGINT Unique identifier of the granted object associated with this mapping. See ki_objects.

object_name VARCHAR(256) Name of the granted object associated with this mapping.

rls VARCHAR(256) Row-level security expression granted in this mapping (for table & view object grants).

cls VARCHAR(256) Column-level security list of columns granted in this mapping (for table & view object grants).

ki_objects

Each record in ki_objects represents a single table, view, or SQL procedure within the database, including those in the virtual catalog tables.

Column Name Column Type Description

oid BIGINT Unique identifier associated with this object.

object_name VARCHAR(256) Name of this object.

schema_oid BIGINT Unique identifier associated with the schema containing this object. See ki_schemas.

schema_name VARCHAR(256) Name of the schema containing this object.

type_id VARCHAR(128) Unique identifier associated with this table's definition (for tables & views).

shard_kind

VARCHAR(1)

Distribution scheme (for regular tables, materialized external tables, and materialized views).

Code	Description
N	Randomly sharded
R	Replicated
S	Sharded, by shard key

persistence

VARCHAR(1)

Whether the object will persist after a database restart.

Code	Description
P	Persisted
T	Temporary

obj_kind

VARCHAR(1)

Object type.

Code	Description
E	External table
H	Partitioned table
I	Materialized view intermediate table
M	Materialized view
P	SQL procedure
R	Regular (local) table
V	Logical view

virtual

VARCHAR(1)

Virtual object type (for virtual objects).

Code	Description
N	Not virtual
F	Filtered view
J	Join view
U	Virtual union

ext_obj_kind

VARCHAR(1)

External object type (for external tables).

Code	Description
N	Not external
L	Logical
M	Materialized

refresh_kind

VARCHAR(1)

Refresh scheme (for refreshable objects).

Code	Description
N	Not refreshable
C	On-change
M	Manual
P	Periodic
Q	On-query
S	On-start

refresh_period DOUBLE Number of seconds between refreshes (for refreshable objects).

column_count INTEGER Number of columns in this table (for tables & views).

has_index INTEGER Whether this object has an index of any kind on any of its columns.

creation_time TIMESTAMP Timestamp of object creation.

last_alter_time TIMESTAMP Timestamp of most recent table definition update (initially, time of creation).

last_update_time TIMESTAMP Timestamp of most recent data update (initially, time of creation).

last_refresh_time TIMESTAMP Timestamp of last data refresh (for refreshable objects; initially, time of creation).

created_by VARCHAR(256) Username of user who created this object.

execute_as VARCHAR(256) Username of user assigned as the execution user of this object (for executable objects).

ttl INTEGER Time-to-live of this object (for objects that can expire).

definition VARCHAR(256) Query that backs this view (for views).

comments VARCHAR(256) User comments associated with this object.

ki_query_active_all

Each record in ki_query_active_all represents the portion of a given query or REST endpoint call being run on a given worker rank at the time this catalog table was queried. It returns general information about the query portion running on the given worker rank. See ki_query_workers for worker-specific details.

Column Name Column Type Description

job_id BIGINT Unique identifier of the system job responsible for the query.

query_id UUID Unique identifier of the query being run.

user_name VARCHAR(64) Name of the user who initiated the query.

resource_group VARCHAR(64) Name of resource group governing this query's resource management.

source_ip VARCHAR(64) IP address of user who initiated the query.

endpoint VARCHAR(64) REST endpoint called to initiate the request (for SQL-based requests).

execution_status

VARCHAR(32)

Status of the query processing.

Status	Description
`none`	Job not yet queued for processing.
`async job done`	Asynchronous job completed successfully.
`cancelled`	Job cancelled.
`cancelling`	Job in the process of being cancelled.
`dispatched`	Job sent to worker ranks for processing.
`done at TOM`	Portion of job completed on worker rank.
`done`	Job completed successfully.
`error`	Job completed with errors.
`initialized`	Persistent job (e.g., PostgreSQL wireline connection) established.
`local cleanup`	Job artifacts being removed before completion.
`running`	Job being processed on worker ranks.

error_message VARCHAR Error message associated with failed query.

start_time TIMESTAMP Timestamp of query processing start.

query_text VARCHAR SQL command issued (for SQL-based requests).

user_data VARCHAR Supplemental information about the request.

sql_step_count INTEGER Step number in processing this SQL command (for SQL-based requests).

refresh_id BIGINT Unique identifier of the materialized view refresh being performed (for materialized view refreshes).

is_mh BOOLEAN Whether the job is a multi-head ingest request.

is_perpetual

BOOLEAN

Whether the job has no fixed end time (e.g., Kafka streaming data feed).

Code	Description
false	Job has a deterministic end time (fixed-size load).
true	Job has no deterministic end time (Kafka feed).

is_cancellable BOOLEAN Whether the job can be cancelled by a user.

is_using_timeout BOOLEAN Whether the job can time out or not.

source_rank INTEGER Number of the rank associated with this job.

ki_query_history

Each record in ki_query_history represents a single query or REST endpoint call made to the database. It will include all SQL queries, but not direct calls to DML-related endpoints, like:

Column Name Column Type Description

job_id BIGINT Unique identifier of the system job responsible for the query.

query_id UUID Unique identifier of the query.

user_name VARCHAR(64) Name of the user who initiated the query.

endpoint VARCHAR(64) REST endpoint called to initiate the request (for SQL-based requests).

execution_status

VARCHAR(32)

Final status of the query processing.

Status	Description
`async job done`	Asynchronous job completed successfully.
`cancelled`	Job cancelled.
`done`	Job completed successfully.
`error`	Job completed with errors.

error_message VARCHAR Error message associated with failed query.

query_text VARCHAR SQL command issued (for SQL-based requests).

user_data VARCHAR Supplemental information about the request.

start_time TIMESTAMP Timestamp of query processing start.

stop_time TIMESTAMP Timestamp of query processing stop.

sql_step_count INTEGER Step number in processing this SQL command (for SQL-based requests).

refresh_id BIGINT Unique identifier of the materialized view refresh performed (for materialized view refreshes).

resource_group VARCHAR(64) Name of resource group that governed this query's resource management.

source_ip VARCHAR(64) IP address of user who initiated the query.

ki_query_span_metrics_all

Each record in ki_query_span_metrics_all represents an independent operation that was processed on a given rank in order to fulfill a given query request, along with the statistics collected for that operation. This table will have one entry for each rank involved in processing a given operation and the statistics collected for that rank.

Column Name	Column Type	Description
`query_id`	`UUID`	Unique identifier of the query request. See ki_query_active_all & ki_query_history.
`span_id`	`UNSIGNED` `BIGINT`	Unique identifier of the metric span for this query.
`parent_span_id`	`UNSIGNED` `BIGINT`	Not used.
`operator`	`VARCHAR(32)`	Internal name for operation performed. Many operators will roughly match the name of the REST endpoints they serve.
`sql_step`	`INTEGER`	Step number in processing this SQL command (for SQL-based requests).
`metric_data`	`JSON`	JSON block of collected metrics.
`start_time`	`TIMESTAMP`	Timestamp of operation start on the associated rank.
`stop_time`	`TIMESTAMP`	Timestamp of operation stop on the associtated rank.
`source_rank`	`INTEGER`	Number of the rank on which this portion of the operation was processed and metrics collected.

ki_query_workers

Each record in ki_query_workers represents the portion of a given query or REST endpoint call being run on a given worker rank at the time this catalog table was queried. It returns worker-specific information about the query portion running on the given worker rank. See ki_query_active_all for general query processing details.

Column Name	Column Type	Description
`job_id`	`BIGINT`	Unique identifier of the system job responsible for the query.
`worker_id`	`BIGINT`	Unique identifier of the worker job responsible for its portion of the query.
`type`	`VARCHAR(64)`	Type of worker process handling this portion of the query.
`internal_name`	`VARCHAR(64)`	Internal name of worker process handling this portion of the query, if present.
`task_info`	`VARCHAR(64)`	Worker task-specific information.
`thread_pool`	`VARCHAR(16)`	Type of thread pool being used to service this portion of the query.
`status`	`VARCHAR(16)`	Status of the processing of this portion of the query--one of: `cancelled` `completed` `failed` `paused` `running`
`blockers`	`VARCHAR(128)`	Any dependencies that may be blocking the execution of this portion of the query.
`sql_step`	`INTEGER`	Step number in processing this SQL command (for SQL-based requests).
`final_msg_set`	`BOOLEAN`	Whether the final status has been recorded for this portion of the query.
`running_task_count`	`INTEGER`	Number of running tasks relating to this portion of the query.
`queued_task_count`	`INTEGER`	Number of queued tasks relating to this portion of the query.
`pending_async_tasks`	`INTEGER`	Number of asynchronous tasks not yet complete for this portion of the query.
`elapsed_time_ms`	`BIGINT`	Number of milliseconds that have passed since processing began on this portion of the query.
`exception`	`VARCHAR(256)`	Any exception thrown during processing this portion of the query.
`source_rank`	`INTEGER`	Number of the rank on which this portion of the query is being processed.

ki_rag_embeddings

Each record in ki_rag_embeddings represents a table or sample definition of a SQL-GPT context and its associated embeddings.

Column Name Column Type Description

obj_id VARCHAR Unique identifier associated with this SQL-GPT context table or sample definition.

context_name VARCHAR(256) Name of the SQL-GPT context containing this definition.

obj_type

VARCHAR(8)

Type of SQL-GPT context definition this is.

Type	Description
`SAMPLE`	Sample definition
`TABLE`	Table definition

embed_text VARCHAR Semicolon-separated compaction of the context definition fields.

n_tokens INTEGER Number of tokens used in generating the embedding for this context definition.

embedding VECTOR(2048) Embedding generated for this context definition.

embed_model VARCHAR(256) Name of the model used to generate the embedding.

ki_role_members

Each record in ki_role_members represents the mapping of a role to a user or role that is a member of that role.

Column Name	Column Type	Description
`role_oid`	`BIGINT`	Unique identifier of the parent role associated with this mapping. See ki_users_and_roles.
`role_name`	`VARCHAR(256)`	Name of the parent role associated with this mapping.
`member_oid`	`BIGINT`	Unique identifier of the user or role that has (is a member of) the parent role. See ki_users_and_roles.
`member_name`	`VARCHAR(256)`	Name of the user or role that has (is a member of) the parent role.
`grantor_oid`	`BIGINT`	Not used.
`admin_option`	`INTEGER`	Whether the member user/role can grant membership to the parent role to other users/roles; effectively, whether the member user/role has system_admin or system_user_admin permission.

ki_schemas

Each record in ki_schemas represents a schema in the database.

Column Name	Column Type	Description
`oid`	`BIGINT`	Unique identifier associated with this schema.
`schema_name`	`VARCHAR(256)`	Name of this schema.
`created_by`	`VARCHAR(256)`	Username of user who created this schema.
`creation_time`	`TIMESTAMP`	Timestamp of schema creation.
`last_alter_time`	`TIMESTAMP`	Timestamp of most recent schema modification (initially, time of creation).
`comments`	`VARCHAR(256)`	Not used.

ki_tiered_objects

Each record in ki_tiered_objects represents a column of data from a given table, located within a given rank & tier. For instance, the oid column of the ki_objects table will likely have one entry for each combination of rank and tier in the database; on a database with two worker ranks, the column will have one entry for the RAM tier and one for the PERSIST tier on each of the two ranks, for a total of four entries for that single table column of data.

Column Name	Column Type	Description
`size`	`BIGINT`	Total number of bytes used by this object in this tier.
`id`	`VARCHAR(256)`	Unique identifier of this tiered object; for tables, containing the tiered column & table name.
`priority`	`INTEGER`	Eviction priority for this object in this tier.
`tier`	`INTEGER`	Name of the tier associated with this object.
`evictable`	`BOOLEAN`	Whether this tiered object is evictable from this tier.
`locked`	`BOOLEAN`	Whether this tiered object is being processed by the tier management system.
`pin_count`	`INTEGER`	Number of operations currently using this tiered object.
`owner_resource_group`	`VARCHAR(128)`	Name of resource group governing this tiered object's tier position and eviction strategy.
`source_rank`	`INTEGER`	Number of the rank associated with this tiered object.
`outer_object`	`VARCHAR(256)`	Not used.

ki_users_and_roles

Each record in ki_users_and_roles represents either a user or role.

Column Name Column Type Description

oid BIGINT Unique identifier associated with this user or role.

name VARCHAR(256) Name of this user or role.

can_login

INTEGER

Whether this user/role can log in to the database; effectively, whether this is a user or role.

Code	Description
0	Can't log in (role)
1	Can log in (user)

is_external

INTEGER

Whether this user is externally authenticated or not.

Code	Description
0	Internal user
1	External user

is_superuser

INTEGER

Whether this user/role has system_admin permission.

Code	Description
0	Doesn't have system admin
1	Has system admin

can_create_role

INTEGER

Whether this user/role can create other users/roles; effectively, whether this user/role has system_admin or system_user_admin permission.

Code	Description
0	Can't create roles
1	Can create roles

can_create_schema

INTEGER

Whether this user/role can create schemas; effectively, whether this user/role has system_admin or system_write permission.

Code	Description
0	Can't create schemas
1	Can create schemas

can_bypass_rls

INTEGER

Whether this user/role can bypass row-level security; effectively, whether this user/role has system_admin permission.

Code	Description
0	Can't bypass row-level security
1	Can bypass row-level security

resource_group INTEGER resource group assigned to this user/role.

comments VARCHAR(256) Not used.