Skip to main content
Version: 24.3.x

ALTER

Update a table’s definition or schema.

Syntax 
ALTER TABLE <table_name>
{ ADD PRIMARY KEY ( <column_name> [, ... ] ) 
  | DROP PRIMARY KEY
  | ADD COLUMNS ( <column_name1> <data_type> [, ... ] )
  | DROP COLUMN <column_name> 
  | { ALTER | MODIFY | CHANGE } COLUMN <old_name> <new_name> <data_type>
  | MODIFY COLUMN <column_name>
      SET MASKING POLICY <function_name> ( <column_name> [, ... ] ) 
      | UNSET MASKING POLICY <function_name> 
  | { ADD | DROP } ROW ACCESS POLICY <function_name> ( <column_name> [, ... ] )
  | LOCALSORT BY ( <column_name1>, <column_name2>, ... )
  | REFRESH METADATA
      [ FOR PARTITIONS ( <partition_name> = '<value>') ]
      [ { AVOID | AUTO } PROMOTION ]
      [ { FORCE | LAZY } UPDATE ]
      [ { MAINTAIN | DELETE } WHEN MISSING ]
  | FORGET METADATA
  | CREATE AGGREGATE REFLECTION <reflection_name> USING 
      { DIMENSIONS ( <column_name1>, <column_name2>, ... ) 
      | MEASURES ( <column_name1> ( <aggregation_type>, <column_name2> <aggregation_type> , ... ) ) 
      | DIMENSIONS ( <column_name1>, <column_name2>, ... ) MEASURES ( <column_name1> ( <aggregation_type>, <column_name2> <aggregation_type> , ... ) ) }
      [ PARTITION BY ( [<column_name>|<partition_transform>] [ , ... ] ) ]
      [ LOCALSORT BY ( <column_name1>, <column_name2>, ... ) ]
      [ ARROW CACHE ]
  | CREATE EXTERNAL REFLECTION <reflection_name> USING <table_name>
  | CREATE RAW REFLECTION <reflection_name> USING DISPLAY ( <column_name1>, <column_name2>, ...)
      [ PARTITION BY ( [<column_name>|<partition_transform>] [ , ... ] ) ]
      [ LOCALSORT BY ( <column_name1>, <column_name2>, ... ) ]
      [ ARROW CACHE ]
  | DROP REFLECTION <reflection_name>
  | ROUTE REFLECTIONS TO { DEFAULT ENGINE | ENGINE { <engine_name> | <engine_uuid> } }
  | { ADD | DROP } PARTITION FIELD { <column_name> | <partition_transform> } 
}

Parameters​

ALTER TABLE <table_name> String

The name of the table that you want to alter.

ADD PRIMARY KEY ( <column_name> [, ... ] ) String

Specifies to use one or more existing columns in a table as the primary key of that table. Uniqueness of the values in a primary key is not enforced. Primary keys are of use in giving the query planner hints for selecting join plans.

DROP PRIMARY KEY String

Removes a table's primary key. The columns that make up the primary key remain in the table.

ADD COLUMNS ( <column_name1> <data_type> [, ... ] ) String

Creates one or more columns that have the specified names, data types, and character limits. The size is an attribute of the data type.

These are the supported primitive types:

  • BOOLEAN
  • VARBINARY
  • DATE
  • FLOAT
  • DECIMAL
  • DOUBLE
  • INTERVAL
  • INT
  • BIGINT
  • TIME
  • TIMESTAMP
  • VARCHAR (The length is always 65536 bytes. If a length is specified, it is ignored.)

You can define complex types by using either of these two sets of syntax:

Set 1

  • struct_type: ROW( name primitive_or_complex_type, .. )
  • list_type: ARRAY(primitive_or_complex_type)

Examples:

ROW(innerfield INT, anotherinnerfield DOUBLE)
ARRAY(INT)
ROW(innerfield ARRAY(INT))
ARRAY(ROW(innerfield INT))

Set 2

  • struct_type: STRUCT <name : primitive_or_complex_type, ... >
  • list_type: { LIST | ARRAY } < primitive_or_complex_type >

Examples:

STRUCT<innerfield : INT, anotherinnerfield : DOUBLE>
LIST<INT>
ARRAY<INT>
STRUCT<innerfield : LIST<INT>>
LIST<STRUCT<innerfield : INT>>

DROP COLUMN <column_name> String

Drops the specified column. This action cannot be undone.

{ ALTER | MODIFY | CHANGE } COLUMN <old_name> <new_name> <data_type> String

Changes the data type for a column, and gives you the option to rename the column.

note

You can rename only columns of tables that are in scratch spaces and columns of Iceberg tables. If you do not want to rename the column, set <new_name> to <old_name>.

Only three types of changes to primitive types are allowed:

  • int to long
  • float to double
  • decimal(P, S) to decimal(P', S), if you are widening the precision

You can alter columns that use complex types by using either of these two sets of syntax:

Set 1

  • struct_type: ROW( name primitive_or_complex_type, .. )
  • list_type: ARRAY(primitive_or_complex_type)

Examples:

ROW(innerfield INT, anotherinnerfield DOUBLE)
ARRAY(INT)
ROW(innerfield ARRAY(INT))
ARRAY(ROW(innerfield INT))

Set 2

  • struct_type: STRUCT <name : primitive_or_complex_type, ...>
  • list_type: { LIST | ARRAY } < primitive_or_complex_type >

Examples:

STRUCT<innerfield : INT, anotherinnerfield : DOUBLE>
LIST<INT>
ARRAY<INT>
STRUCT<innerfield : LIST<INT>>
LIST<STRUCT<innerfield : INT>>

MODIFY COLUMN <column_name> String

Specifies the column on which to either set a masking policy or unset one.

SET MASKING POLICY <function_name> ( <column_name> [, ... ] ) String

Sets a policy for masking data. The UDF serving as the masking policy must accept and return the same data type as the column it is masking. If a UDF with the given name does not exist, then the affected table will not be reachable until the policy is dropped or a UDF created.

UNSET MASKING POLICY <function_name> String

Removes a policy for masking data.

{ ADD | DROP } ROW ACCESS POLICY <function_name> ( <column_name> [, ... ] ) String

Adds a row-access policy to the table, or removes one from the table. The UDF serving as the row-access policy must return BOOLEAN. If the UDF specified in the policy does not exist, then the Dremio engine gives an error when attempting to query the table.

LOCALSORT BY ( <column_name1>, <column_name2>, ... ) String   Optional

The columns on which to sort new data that is added to the table after the ALTER TABLE command is run.

REFRESH METADATA

Refreshes the metadata associated with the specified table. You can choose to either do a full or partial metadata refresh. A full refresh will update the metadata for the entire table. A partial refresh will enable you to only update specified partitions. Optional clauses are available for refreshing a table's metadata. If you choose to use two or more of these clauses, they must be entered in a specified priority order:

  1. Promotion option: either AVOID PROMOTION or AUTO PROMOTION
  2. Update option: either FORCE UPDATE or LAZY UPDATE
  3. Missing option: either MAINTAIN WHEN MISSING or DELETE WHEN MISSING
note

If the table is not partitioned, you will receive an error when you attempt to refresh the metadata using the FOR PARTITIONS clause.

FOR PARTITIONS ( <partition_name> = '<value>' ) String   Optional

Use this clause to do a partial refresh of the table's metadata.

  • The <partition_name> identifies the name of the partition to be refreshed.
  • The <value> identifies the specific partition that should be refreshed. Must be contained in single quotes.

{ AVOID | AUTO } PROMOTION Optional

Clauses that determine whether files and folders are promoted to datasets when you run a query.

The AVOID PROMOTION prevents queries from promoting files/folders to datasets. The AUTO PROMOTION allows queries to promote files/folders to datasets. This is the default option when you do not include a promotion clause.

{ FORCE | LAZY } UPDATE Optional

Clauses that determine whether metadata is updated when you run a query.

The FORCE UPDATE forces a full update of metadata. The LAZY UPDATE does not perform a full update of metadata. This is the default option when you do not include an update clause.

{ MAINTAIN | DELETE } WHEN MISSING Optional

Clauses that determine how missing metadata is handled when you run a query.

The MAINTAIN WHEN MISSING prevents missing metadata from being deleted during refresh. DELETE WHEN MISSING deletes missing metadata during refresh. This is the default option when you do not include a clause.

FORGET METADATA

Deletes the metadata information stored in Dremio for the specified table until the next metadata refresh. The dataset can still be queried using SQL.

SET TBLPROPERTIES ('property_name' = 'property_value', ...)

Sets one or more table properties for controlling the behavior of the table. See Properties of Apache Iceberg Tables for a list of the supported properties.

UNSET TBLPROPERTIES ('property_name' = 'property_value', ...)

Removes one or more table properties that control the behavior of the table. See Properties of Apache Iceberg Tables for a list of the supported properties.

CREATE AGGREGATE REFLECTION <reflection_name> String

Creates an aggregation reflection with the specified name.

DIMENSIONS ( <column_name1>, <column_name2>, ... ) String

The columns to include as dimensions in the reflection.

MEASURES ( <column_name1> ( <aggregation_type> ), <column_name2> ( <aggregation_type ), ... ) String

The columns to include as measures in the reflection, and the type of aggregation to perform on them. The possible types are COUNT, MIN, MAX, SUM, and APPROXIMATE COUNT DISTINCT.

PARTITION BY ( [<column_name>|<partition_transform>] [ , ... ] ) String   Optional

The columns or the partition transforms on which to partition the data horizontally in the reflection. If a column is listed as a partition column, it cannot also be listed as a sort column for the same reflection. Each column specified as a partition column or used in transform must also be listed as a dimension column.

Type of Partition TransformDescription
identity(<column_name>)Creates one partition per value. This is the default transform. The column can use any supported data type.
year(<column_name>)Partitions by year. The column must use the DATE or TIMESTAMP data type.
month(<column_name>)Partitions by month. The column must use the DATE or TIMESTAMP data type.
day(<column_name>)Partitions on the equivalent of dateint. The column must use the DATE or TIMESTAMP data type.
hour(<column_name>)Partitions on the equivalent of dateint and hour. The column must use the TIMESTAMP data type.
bucket(<integer>, <column_name>)Partitions data into the number of partitions specified by an integer. For example, if the integer value N is specified, the data is partitioned into N, or (0 to (N-1)), partitions. The partition in which an individual row is stored is determined by hashing the column value and then calculating <hash_value> mod N. If the result is 0, the row is placed in partition 0; if the result is 1, the row is placed in partition 1; and so on.

The column can use the DECIMAL, INT, BIGINT, VARCHAR, VARBINARY, DATE, or TIMESTAMP data type.
truncate(<integer>, <column_name>)If the specified column uses the string data type, truncates strings to a maximum of the number of characters specified by an integer. For example, suppose the specified transform is truncate(1, stateUS). A value of CA is truncated to C, and the row is placed in partition C. A value of CO is also truncated to C, and the row is also placed in partition C.

If the specified column uses the integer or long data type, truncates column values in the following way: For any truncate(L, col), truncates the column value to the biggest multiple of L that is smaller than the column value. For example, suppose the specified transform is truncate(10, intColumn). A value of 1 is truncated to 0 and the row is placed in the partition 0. A value of 247 is truncated to 240 and the row is placed in partition 240. If the transform is truncate(3, intColumn), a value of 13 is truncated to 12 and the row is placed in partition 12. A value of 255 is not truncated, because it is divisble by 3, and the row is placed in partition 255.

The column can use the DECIMAL, INT, BIGINT, VARCHAR, or VARBINARY data type.

Note: The truncate transform does not change column values. It uses column values to calculate the correct partitions in which to place rows.

LOCALSORT BY ( <column_name1>, <column_name2>, ... ) String   Optional

The columns on which to sort the data that is in the reflection.

ARROW CACHE Optional

Specifies that you want Dremio to convert data from your reflection’s Parquet files to the Apache Arrow format when copying that data to executor nodes. Normally, Dremio copies data as-is from the Parquet files to caches on executor nodes, which are nodes that carry out the query plans devised by the query optimizer. Enabling this option can improve query performance even more. However, data in the Apache Arrow format requires more space on the executor nodes than data in the default format. You can use this option with Amazon S3.

CREATE EXTERNAL REFLECTION <reflection_name> String

Creates an external reflection with the specified name.

USING <table_name> String

The path of the derived table to use with the external reflection.

CREATE RAW REFLECTION <reflection_name> String

Creates a raw reflection with the specified name.

USING DISPLAY ( <column_name1>, <column_name2>, ... ) String

The columns to include in the reflection.

PARTITION BY ( [<column_name>|<partition_transform>] [ , ... ] ) String   Optional

The columns or the partition transforms on which to partition the data horizontally in the reflection. If a column is listed as a partition column, it cannot also be listed as a sort column for the same reflection. Each column specified as a partition column or used in transform must also be listed as a display column.

Type of Partition TransformDescription
identity(<column_name>)Creates one partition per value. This is the default transform. The column can use any supported data type.
year(<column_name>)Partitions by year. The column must use the DATE or TIMESTAMP data type.
month(<column_name>)Partitions by month. The column must use the DATE or TIMESTAMP data type.
day(<column_name>)Partitions on the equivalent of dateint. The column must use the DATE or TIMESTAMP data type.
hour(<column_name>)Partitions on the equivalent of dateint and hour. The column must use the TIMESTAMP data type.
bucket(<integer>, <column_name>)Partitions data into the number of partitions specified by an integer. For example, if the integer value N is specified, the data is partitioned into N, or (0 to (N-1)), partitions. The partition in which an individual row is stored is determined by hashing the column value and then calculating <hash_value> mod N. If the result is 0, the row is placed in partition 0; if the result is 1, the row is placed in partition 1; and so on.

The column can use the DECIMAL, INT, BIGINT, VARCHAR, VARBINARY, DATE, or TIMESTAMP data type.
truncate(<integer>, <column_name>)If the specified column uses the string data type, truncates strings to a maximum of the number of characters specified by an integer. For example, suppose the specified transform is truncate(1, stateUS). A value of CA is truncated to C, and the row is placed in partition C. A value of CO is also truncated to C, and the row is also placed in partition C.

If the specified column uses the integer or long data type, truncates column values in the following way: For any truncate(L, col), truncates the column value to the biggest multiple of L that is smaller than the column value. For example, suppose the specified transform is truncate(10, intColumn). A value of 1 is truncated to 0 and the row is placed in the partition 0. A value of 247 is truncated to 240 and the row is placed in partition 240. If the transform is truncate(3, intColumn), a value of 13 is truncated to 12 and the row is placed in partition 12. A value of 255 is not truncated, because it is divisble by 3, and the row is placed in partition 255.

The column can use the DECIMAL, INT, BIGINT, VARCHAR, or VARBINARY data type.

Note: The truncate transform does not change column values. It uses column values to calculate the correct partitions in which to place rows.

LOCALSORT BY ( <column_name1>, <column_name2>, ... ) String   Optional

The columns on which to sort the data that is in the reflection.

ARROW CACHE Optional

Specifies that you want Dremio to convert data from your reflection’s Parquet files to the Apache Arrow format when copying that data to executor nodes. Normally, Dremio copies data as-is from the Parquet files to caches on executor nodes, which are nodes that carry out the query plans devised by the query optimizer. Enabling this option can improve query performance even more. However, data in the Apache Arrow format requires more space on the executor nodes than data in the default format. You can use this option with Amazon S3.

DROP REFLECTION String   Optional

Drops the specified reflection defined on the table.

ROUTE REFLECTIONS TO { DEFAULT ENGINE | ENGINE { <engine_name> | <engine_uuid> } } String

Specify the engine on which to run jobs that create and refresh reflections that are defined on the table. Use either the name or UUID of the engine. If neither a name nor a UUID is specified, the default engine is used. You can also directly specify the default engine.

{ ADD | DROP } PARTITION FIELD { <column_name> | <partition_transform> } String   Optional

ADD partitions the table data on the values in a single column or by using one of the partition-transformation functions. DROP drops the partition definition. These are the partition-transformation functions:

TransformDescription
identity( <col> )Explicitly specified identity transform
year( <col> )Partition by year. The column uses either the TIMESTAMP or DATE data type.
month( <ts_col> )Partition by month. The column uses either the TIMESTAMP or DATE data type.
day( <ts_col> )Partition by day. The column uses either the TIMESTAMP or DATE data type.
hour( <ts_col> )Partition by hour. The column uses the TIMESTAMP data type.
bucket( <count>, <col> )Partition by hashed value into <count> buckets
truncate( <length>, <col> )Partition by truncated value.
  • Strings are truncated to the specified length.
  • Integer and biginteger values are truncated to bins.
    Example: truncate(10, i) produces 0, 10, 20, and so on.

Examples​

Add a column 
ALTER TABLE services
  ADD COLUMNS (county varchar)
Modify a column 
ALTER TABLE services
  MODIFY COLUMN tip_amount tip_amount DECIMAL
Modify a column for a struct table 
ALTER TABLE struct_type
  MODIFY COLUMN a  a struct<x: varchar, y: bigint>
Rename a column 
ALTER TABLE services
  MODIFY COLUMN tip_amount gratuity_amount DECIMAL
Refresh all the metadata for a table 
ALTER TABLE services
  REFRESH METADATA
Refresh all the metadata for a table using optional clauses 
ALTER TABLE services
  REFRESH METADATA
    AUTO PROMOTION
    LAZY UPDATE
    MAINTAIN WHEN MISSING
Refresh the metadata for a single partition 
ALTER TABLE Samples."samples.dremio.com"."zips.json"  
  REFRESH METADATA FOR PARTITIONS (state = 'TX')
Refresh the metadata for a single partition using optional clauses 
ALTER TABLE Samples."samples.dremio.com"."zips.json"  
  REFRESH METADATA FOR PARTITIONS (state = 'TX')
    AUTO PROMOTION
    LAZY UPDATE
    MAINTAIN WHEN MISSING
Forget the metadata for a table 
ALTER TABLE Samples."samples.dremio.com"."zips.json"
  FORGET METADATA
Create a raw reflection that sorts customers by last name and partitions them by country 
ALTER TABLE "@user1"."customers"
  CREATE RAW REFLECTION customers_by_country USING DISPLAY (id,lastName,firstName,address,country)
    PARTITION BY (country)
    LOCALSORT BY (lastName)
Create an aggregate reflection that counts the cities per state and sorts by state 
ALTER TABLE Samples."samples.dremio.com"."zips.json"
  CREATE AGGREGATE REFLECTION per_state
  USING
  DIMENSIONS (state)
  MEASURES (city (COUNT))
  LOCALSORT BY (state)
Routing Reflections 
ALTER TABLE "Table 1" ROUTE REFLECTIONS TO ENGINE "Engine 1"
ALTER TABLE "View 1" ROUTE REFLECTIONS TO ENGINE "Engine 1"
ALTER TABLE "View 3" ROUTE REFLECTIONS TO ENGINE "Engine 1"
Adding a column for an Apache Iceberg Table 
ALTER TABLE myTable ADD COLUMNS (address VARCHAR)
Changing the data type of a column to BIGINT for an Apache Iceberg Table 
ALTER TABLE myTable ALTER COLUMN id id BIGINT
Set a column-masking policy that takes multiple columns 
ALTER TABLE e.employees
    MODIFY COLUMN ssn_col
    SET MASKING POLICY protect_ssn (ssn_col, region)
Unset a column-masking policy 
ALTER TABLE e.employees
    MODIFY COLUMN ssn_col
    UNSET MASKING POLICY protect_ssn
Add a row-access policy to a table 
ALTER TABLE e.employees
  ADD ROW ACCESS POLICY state_policy ( state_col )
Remove a row-access policy from a table 
ALTER TABLE employees
  DROP ROW ACCESS POLICY protect_ssn (ssn_col, region)