Link Search Menu Expand Document

Release notes archive

We provide an archive of release notes for your historical reference.

DB version 3.31

February 2024

Enhancements, changes and new integrations

Simplified table protobuf representation

Unique constraints in tables will be blocked for new accounts.

Renamed spilled metrics columns

The columns spilled_bytes_uncompressed and spilled_bytes_compressed of information_schema.query_history have been replaced by a single column spilled_bytes. It contains the amount of data that was spilled to disk temporarily while executing the query.

New requirements updated for EXPLAIN

For EXPLAIN queries, we now allow only one of the following options at the same time: ALL, LOGICAL, PHYSICAL, ANALYZE.EXPLAIN (ALL) now returns the plans in multiple rows instead of multiple columns.

Range violation implement for import of parquet INT columns into PDGATE columns

Reading of Parquet/ORC integer columns will now not be allowed if the external table specifies the types of those columns to be one of the new DATE, TIMESTAMP, TIMESTAMPTZ types.

Resolved issues

Fixed a bug where floating point values -0.0 and +0.0, as well as -nan and +nan were not considered equal in distributed queries.

Updated error log for upload failure for clarity

TRY_CAST from TEXT to NUMERIC now works as expected: if the value cannot be parsed as NUMERIC it produces null.

DB version 3.30

November 2023

New features

New comparison operators

New comparison operators IS DISTINCT FROM and IS NOT DISTINCT FROM have been added.

Enhancements, changes and new integrations

Support for nullable arrays

Support has been added to allow the ANY_MATCH lambda function to work with nullable arrays.

Resolved issues

  • Indirectly granted privileges have been removed from the information_schema.object_privileges view.

  • Fixed an issue where ARRAY_FIRST and ARRAY_FIRST_INDEX returned an error if the given input was nullable.

DB version 3.30

November 2023

New features

New comparison operators

New comparison operators IS DISTINCT FROM and IS NOT DISTINCT FROM have been added.

Enhancements, changes and new integrations

Support for nullable arrays

Support has been added to allow the ANY_MATCH lambda function to work with nullable arrays.

Resolved issues

  • Indirectly granted privileges have been removed from the information_schema.object_privileges view.

  • Fixed an issue where ARRAY_FIRST and ARRAY_FIRST_INDEX returned an error if the given input was nullable.

DB version 3.29

October 2023

New features

EXPLAIN ANALYZE now available for detailed query metrics

You can now use the EXPLAIN command to execute EXPLAIN (ANALYZE) <select statement> and get detailed metrics about how much time is spent on each operator in the query plan, and how much data is processed. The query plan shown there is the physical query plan, which you can inspect using EXPLAIN (PHYSICAL) <select statement> without executing the query. It shows how query processing is distributed over the nodes of an engine.

Enhancements, changes and new integrations

PERCENTILE_CONT and PERCENTILE_DISC now return Postgres-compliant results

PERCENTILE_CONT for decimal input now returns DOUBLE PRECISION instead of NUMERIC data type.

Virtual column ‘source_file_timestamp’ uses new data type

The virtual column source_file_timstamp has been migrated from the data type TIMESTAMP (legacy timestamp type without time zone) to the type TIMESTAMPTZ (new timestamp type with time zone).

Despite the increased resolution, the data is still in second precision as AWS S3 provides them only as unix seconds.

Use source_file_timestamp - NOW() instead of DATE_DIFF('second', source_file_timestamp, NOW())

New function added

A new alias ARRAY_TO_STRING has been added to function ARRAY_JOIN.

DB version 3.28

September 2023

Resolved issues

  • IN expressions with scalar arguments now return Postgres-compliant results if there are NULLs in the IN list.

  • information_schema.running_queries returns ID of a user that issued the running query, not the current user.

  • Update error message to explain upper case behavior

DB version 3.27

September 2023

New features

  • Add URL_DECODE and URL_ENCODE SQL functions
  • URL_ENCODE percent-encodes all non unreserved characters; for example, SELECT CONCAT('https://www.firebolt.io/?', URL_ENCODE('example_id=1&hl=en')); returns: https://www.firebolt.io/?example_id%3D1%26hl%3Den

  • URL_DECODE decodes percent-encoded characters; for example, SELECT URL_DECODE('https://www.firebolt.io/?example_id%3D1%26hl%3Den'); returns: https://www.firebolt.io/?example_id=1&hl=en

  • New GENERATE_SERIES function

Support has been added for the GENERATE_SERIES function to generate a list of values based on the start, stop, and optional configurable increment step.

Enhancements, changes, and new integrations

  • Updates to CSV ingestion TRUE and FALSE are now accepted as valid BOOLEAN values when ingesting CSV files.

DB version 3.26

September 2023

New features

  • New setting to limit rows in result

    A new setting max_result_rows controls the limit of rows in result sets. The default value of the max_result_rows setting is 0.

Enhancements, changes and new integrations

  • Improved support for interval arithmetic

    You can now use the expression: date_time + INTERVAL * d where date_time is a constant or column reference of type DATE, TIMESTAMP, PGDATE, TIMESTAMPNTZ, or TIMESTAMPTZ, and d is a constant or column reference of type DOUBLE PRECISION. The effect is that the INTERVAL is scaled by d, and the resulting INTERVAL is added to date_time. For example: INTERVAL '1 day' * 3 -> INTERVAL '3 days'

  • Added support for BYTEA inputs in the LENGTH function

The LENGTH function now accepts an input of BYTEA type. The function returns the number of bytes of an input byte array. For behaviors of converting TEXT strings to BYTEA, please refer to BYTEA data type.

Resolved issues

  • Fixed an issue where query progress was not reflected in the information_schema.running_queries table.

  • Significantly increased performance of the COPY TO function.

DB version 3.25

August 2023

Enhancements, changes and new integrations

  • Deprecation of catalog metadata schema

    Support for the catalog schema is being phased out in favor of information_schema views. To ensure a smooth transition, please update code to use information_schema - for example, any query currently reading from catalog.query_history or catalog.running_queries should be modified to query from information_schema.query_history or information_schema.running_queries. Please note, all column names in information_schema views are lowercase - for example, START_TIME from catalog.query_history is now named start_time in the information_schema.query_history view.

  • Information schema updated

    Added columns to the information_schema.databases view:

    • compressed_size
    • uncompressed_size
    • description

    Added columns to the information_schema.engines view:

    • engine_type
    • auto_stop
    • url
    • warmup

Resolved issues

  • Fixed an issue preventing certain array type columns being imported from a Parquet file.

  • Fixed an issue causing inf and nan values to be incorrectly encoded as NULL values.

DB version 3.24

July 2023

New features

  • Added support for functions HLL_COUNT_BUILD, HLL_COUNT_EXTRACT and HLL_COUNT_MERGE_PARTIAL

    HLL_COUNT_BUILD uses the HLL++ algorithm and allows you to control the set sketch size precision, aggregating input values to an HLL++ sketch represented as the BYTEA data type. Later individual sketches can be merged to a single sketch using the aggregate function HLL_COUNT_MERGE_PARTIAL, or the estimated cardinality extracted (to get the final estimated distinct count value) using the HLL_COUNT_EXTRACT scalar function.

  • Added support for service accounts

    Service account users (Beta) are now available. Read more about using service accounts here.

  • Added support for new function PARAM

    Use the new PARAM function to reference values of query parameters.

  • Added support for VACUUM command

    The VACUUM command is now generally available, for use in production workflows.

Enhancements, changes and new integrations

  • Change to SUBSTRING function

    The SUBSTRING function with updated behavior is now available and is no longer an alias for the SUBSTR (legacy) function. With the SUBSTRING function, negative offsets are treated as offset 1, thus starting at the beginning of the input string. With the SUBSTR (legacy) function, negative offset values indicate an offset from the end of the input string. With the new SUBSTRING function, for index values less than 1, the length is decreased by the difference between 1 and the index value. Negative length values are no longer allowed, and indexing is now 1-based, rather than 0-based as with the SUBSTR (legacy) function.

  • Added UTF-8 validation for text fields

    All text fields (both literals and those from an external source) must be UTF-8 encoded to pass validation. To store strings that are not UTF-8 encoded, use the BYTEA data type as an alternative.

  • Update to the relationship of floating point NaN to other numbers

    The value of floating point NaN is always the largest when compared to other numeric values, for example: -∞ < any number < ∞ < NaN.

  • Unsupported functions behavior

    Functions not found in the Firebolt SQL reference will be blocked. If you have scripts using these functions, please work with your Customer Success team to implement supported functions.

Resolved issues

  • Fixed an issue preventing columns with spaces from being used in the primary index definition.

  • Fixed an issue where adding filters on a partition key could affect query performance.

  • Fixed an issue causing chained UNION/INTERSECT operations to be applied in the wrong order.

  • Significant performance improvements made to window functions with PARTITION BY in the frame specification.

DB version 3.23

May 2023

New features

  • Added support for functions REGEXP_EXTRACT and REGEXP_EXTRACT_ALL

    Use these functions to extract matching patterns within an input expression. The REGEXP_EXTRACT function extracts the first match only (from the left), REGEXP_EXTRACT_ALL function extracts all the matches.

  • System Engine (Beta)

    Use the new system engine to run metadata-related queries without having to start a separate engine. The system engine is always available in all databases to select and use.

    System engine is currently only available for accounts that have a single region enabled.

Enhancements, changes and new integrations

  • Support for AWS Glue with external tables has been removed.

    Using AWS Glue with external tables has been deprecated. Support for this feature may be added back in the future, once optimizations to supporting architecture have been added.

  • Increasing max session duration is no longer required when setting up AWS roles to access Amazon S3

    Roles are now assumed with a default expiration time of 1 hour, and re-assumed every 30 minutes during ingest so ingests can take longer than 12 hours.

  • Increased validation for CREATE TABLE parameters

    Unknown parameters in a CREATE TABLE statement will now cause the statement to be rejected.

  • Update to non-explicit column names

    Final result column names may be different compared to previous versions; specifically, for any select list item that does not have an explicit alias assigned, the final name of that column may be “?column?”. Please assign explicit aliases to any select list expression which you want to refer to either in the same query or in other queries.

  • Join index improvements

    Starting in version 3.23, manually creating and refreshing join indexes will no longer be necessary - with Auto join indexes, results are now always up to date even if the underlying data changed.

Resolved issues

  • Fixed an issue causing a discrepancy in results of the SHOW CACHE command.

  • Fixed multiple planner bugs where queries would not compile.

  • Improved ingest stability, fixed an issue causing failed ingests on a general purpose engine.

DB version 3.22

April 2023

New features

  • DATE and TIMESTAMP names available for new data types

    A new option enables the use of familiar type names DATE and TIMESTAMP as default for the new expanded date and timestamp data types, with synonyms PGDATE and TIMESTAMPNTZ. TIMESTAMPTZ remains the same as a new type added. If you are a new customer starting on DB version 3.22, these new date and timestamp type names will be enabled by default. For more information, see Date and timestamp data types.

    For existing customers before DB version 3.22:

    To use the new data types, new external and dimension/fact tables must be created and scripts updated to use functions that support these new types. Reingest will be required to recognize new precision.

    • To ingest from an existing table into a new table using the new types, simply cast a column of type DATE to PGDATE and a column of type >TIMESTAMP to TIMESTAMPNTZ.
    • To ingest into a new table using the new types from external data, create an external table with the new types.

    See Date and timestamp (legacy) for information about how to adjust scripts for supported functions.

    For existing customers before DB version 3.22, data must be reingested using the new types and scripts updated to use supported functions before this option is enabled. Please contact your Customer Success team to enable the DATE and TIMESTAMP synonymns for new types once you have reingested and adjusted scripts.

  • Schema changes instantly available on analytics engines

    Running engines no longer require a restart for schema changes to be reflected on analytics engines, and changes to the schema are available instantly. Schema changes include:

    • DROP/CREATE TABLE
    • DROP/CREATE AGGREGATING INDEX
    • DROP/CREATE JOIN INDEX
    • DROP/CREATE VIEW

  • Added support for UPDATE and DELETE commands

    Data manipulation commands UPDATE and DELETE are generally available, for use in production workflows.

  • Added support for NUMERIC data type

    The NUMERIC data type (synonym: DECIMAL) is now generally available, for use in production workflows.

Enhancements, changes, and new integrations

  • Support for OR operator added to all JOIN types

    The OR operator can now be used within the join condition for all types of joins.

  • UI option to explain queries (UI release)

    A new option in the UI supports explaining statements for single and multi-statement scripts.

  • Visual explain temporarily limited to text view

    Text view for visual explain will be the only supported view temporarily while improvements to visual explain are developed. Other views will be added back in the future.

Resolved issues

  • Improved error message when casting between data types.

  • Fixed the NOT operator to work with all expressions.

  • Fixed incorrect source_file_name metadata generated when an S3 bucket included empty files.

  • Fixed an issue preventing the flattening of a list of structs when nested structs are included.

  • Fixed an issue causing preloaded tablets to merge in an improper order.

  • Fixed an issue causing a Sync out error during delete queries.

DB version 3.21.0

March 2023

New features

  • Data manipulation commands now available (Beta release)

    Beta support is now available for data manipulation commands UPDATE and DELETE. A VACUUM has also been added to optimize frequently updated tables.

    For more information and known limitations in the beta release, please see linked documentation.

Enhancements, changes, and new integrations

  • Updates to data types

    Firebolt now uses the following built-in type names in INFORMATION_SCHEMA.COLUMNS and in auto-generated aliases for CAST operations:

    Standard type Synonyms
    TEXT STRING, VARCHAR
    INTEGER INT, INT4
    BIGINT LONG, INT8
    REAL FLOAT, FLOAT4
    DOUBLE PRECISION DOUBLE, FLOAT8
    NUMERIC DECIMAL

  • Parquet, Avro and ORC support added for new data types

    These file types can now be used to ingest new PGDATE, TIMSTAMPNTZ and TIMESTAMPTZ data types. For more information see data type documention.

    Starting in the next version, you will have the option to use the type names DATE and TIMESTAMP instead of new type names PGDATE and TIMESTAMPNTZ, but data must be reingested using the new types before this option is enabled. TIMESTAMPTZ will remain the same, as that is a new type added. See here for instructions to reingest. Please raise any questions or feedback with your Customer Success team.

Resolved issues

  • Fixed an error Cannot parse input: expected <END OF LINE> on ingest of large CSV files.

DB version 3.20.0

March 2023

Enhancements, changes, and new integrations

To use the new data types, new external and dimension/fact tables must be created. Reingest will be required to recognize new precision.

  • To ingest from an existing table into a new table using the new types, simply cast a column of type DATE to PGDATE and a column of type >TIMESTAMP to TIMESTAMPNTZ.
  • To ingest into a new table using the new types from external data, create an external table with the new types.

In a future version, you will have the option to use the type names DATE and TIMESTAMP instead of new type names PGDATE and TIMESTAMPNTZ, but data must be reingested using the new types before this option is enabled. TIMESTAMPTZ will remain the same, as that is a new type added. Please raise any questions or feedback with your Customer Success team.

  • Changed NULL behavior of CONCAT function

    NULL inputs to the CONCAT function are now treated as empty strings, therefore any NULL inputs are ignored. When all inputs are NULL, the result will be an empty string. When using ||, any NULL input still results in a NULL output.

If you are using the CONCAT function on strings with NULL inputs and you don’t want NULL values to be ignored, you will need to use the || function instead.

  • Additional syntax for ARRAY data type names

    Syntax options for defining columns with the ARRAY data type have been updated to include <data-type>[] and <data-type> ARRAY. Array element type is nullable when using the new syntax options.

    For example, the following three queries will create tables with the same demo_array column of type ARRAY of nullable TEXT.

    CREATE DIMENSION TABLE demo1 (
demo_array ARRAY(TEXT NULL)
);
  
CREATE DIMENSION TABLE demo2 (
demo_array TEXT[]
);

CREATE DIMENSION TABLE demo3 (
demo_array TEXT ARRAY
);

    To specify the constraint for an array element to be not nullable, you must then use ARRAY(<data-type> NOT NULL) syntax.

  • Added flag support for REGEXP_LIKE

    The REGEXP_LIKE function now supports an optional <flag> input, to allow additional controls over the regular’s expression matching behavior.

    For example, the i flag causes the regular expression matching in the following query to be case-insensitive. Without this flag, the query would not find a match and would return 0.

    SELECT
    REGEXP_LIKE('ABC', '[a-z]', 'i'); ---> 1

  • Parquet and ORC support added for binary data type

    Binary type data from external Parquet or ORC file types will now be ingested directly with the data type BYTEA. Previously, data were ingested as type TEXT and then converted to data type BYTEA.

  • Export all results from the SQL Workspace (UI release)

    Exporting the entire results section from the SQL Workspace in CSV or JSON format is now supported.

  • Renamed column from information_schema.tables

    Renamed the number_of_segments column from information_schema.tables to number_of_tablets to better reflect Firebolt’s data structure.

  • Added support for empty statements

    Empty statements containing comments only are now supported and will run without error.

    SELECT
    REGEXP_LIKE('ABC', '[a-z]', 'i'); ---> 1

  • Parquet and ORC support added for binary data type

    Binary type data from external Parquet or ORC file types will now be ingested directly with the data type BYTEA. Previously, data were ingested as type TEXT and then converted to data type BYTEA.

  • Added support for empty statements

    Empty statements containing comments only are now supported and will run without error.

  • Export all results from the SQL Workspace

    Exporting the entire results section from the SQL Workspace in CSV or JSON format is now supported.

  • Renamed column from information_schema.tables

    Renamed the number_of_segments column from information_schema.tables to number_of_tablets to better reflect Firebolt’s data structure.

  • Added support for empty statements

    Empty statements containing comments only are now supported and will run without error.

Resolved issues

  • Fixed an issue where AVG and SUM functions performed on large DECIMAL columns produced an error; results now use the same precision and scale as the input type.

DB version 3.19.0

February 2023

New features

  • New date and time data types

    Added support for new date and timestamp data types:

    The new data types use an improved memory layout providing a much higher supported range, now extending from 0001-01-01[ 00:00:00.000000] to 9999-12-31[ 23:59:59.999999]. This change also extends the syntax for specifying intervals used for arithmetic with dates and timestamps. In addition to the previously supported interval syntax, you can now also write interval 'N' unit, where N is a possibly signed integer, and unit is one of year, month, day, hour, minute, or second, matched case-insensitively.

    The previously supported DATE and TIMESTAMP data types are planned for deprecation in the future. New features and functionality will be built to support the new date and timestamp data types, rather than these legacy types.

    To use the new data types, new external and dimension/fact tables must be created. Reingest will be required to recognize new precision.

    • To ingest from an existing table into a new table using the new types, simply cast a column of type DATE to PGDATE and a column of type TIMESTAMP to TIMESTAMPNTZ.
    • To ingest into a new table using the new types from external data, create an external table with the new types.

    The new syntax can break the semantics of existing SQL queries. Previously, the unit part of the expression was treated as a column alias, and now it’s treated as part of the interval literal. For example, if you wrote SELECT DATE '2023-01-10' + interval '42' day;, you would get back a table with one column called day and the value 2023-01-10 00:00:42. Now, you will get a back a table with one column (unspecified name) and the value 2023-02-21 00:00:00.
    If you want to retain the old behavior, use AS, for example: SELECT DATE '2023-01-10' + interval '42' AS day;.

  • New setting for time zone

    New setting time_zone controls the session time zone. The default value of the time_zone setting is UTC.

  • New keyboard shortcuts (UI release)

    Use new keyboard shortcuts in the SQL workspace to cancel a query, or go to a specific line in your script.

    • Cancel a running script with Ctrl + Alt + k for Windows & Linux, or ⌘ + Option + k for Mac
    • Go to a desired line with Ctrl + l for Windows & Linux, or ⌘ + l for Mac

Enhancements, changes, and new integrations

  • Improved join index performance

    Join indexes just got better: profit from their extreme performance benefits without any configuration. Moreover, there is no more need to manually create or refresh – the results are always up to date even if the underlying data changed. With this optimization, we’ve seen real-world, production queries run 200x faster.

    To see how this works, let’s look at an example. Say we have the following query pattern which is run hundreds of times per second with different values for l.player_id and l.date:

    SELECT r.name, SUM(l.score) 
FROM   game_plays as l
JOIN   player_info as r 
ON     l.player_id = r.player_id
WHERE  l.player_id = XXX AND l.date = YYY
GROUP  BY r.name;

    On the first run of this query, the relevant data from the right-hand side table player_info is read and stored in a specialized data structure, which is cached in RAM. This can take tens of seconds if the player_info table is large (e.g., contains millions of rows). However, on subsequent runs of the query pattern, the cached data structure can be reused - so all subsequent queries will only take a few milliseconds (if the left-hand side with potential field restrictions is small, as here).

    Requirements for query optimization

    • The right side of the join in the query must be directly a table. Subselects are not supported.
    • Restrictions on fields from the right side of the join need to be applied in an OUTER SELECT, wrapping the query.
    • Since the join index data structure is cached in RAM, the right side table may not be too large (by default the size of the cache is limited to 20% of the RAM).
    • All types of joins (INNER, LEFT, RIGHT, …) are supported.
    • The right table in the join can be a FACT or DIMENSION table.

  • Improved cache eviction

    Cache eviction process and stability has been improved. Tablet eviction is now managed by a Least Recently Used (LRU) algorithm, which provides smarter eviction and keeps the data that is most likely to be accessed in the engine cache.

  • Added syntax option for setting TYPE options in CREATE EXTERNAL TABLE

    Added the option to set type options for S3 source files at the same level as TYPE is set. Type option can now be defined as in the example below:

    CREATE EXTERNAL TABLE ex_table( ... )
TYPE=(CSV)
ALLOW_COLUMN_MISMATCH=true
...

    as well as with the original syntax:

    CREATE EXTERNAL TABLE ex_table(...)
TYPE=(CSV ALLOW_COLUMN_MISMATCH=true)
...

  • Default DECIMAL scale changed

    The default scale for the DECIMAL data type has been updated from 0 to 9.

DB version 3.17.0

February 2023

New features

  • Added support for GROUP BY ALL

    Instead of explicitly listing all grouping elements in the GROUP BY clause, use GROUP BY ALL to automatically infer them from the SELECT list.

  • New BYTEA data type

    Use the new BYTEA data type to store binary data, like images, other multimedia files, or raw bytes of information.

  • New functions ENCODE and DECODE

    Use these functions with the new BYTEA data type to encode binary data into a SQL expression of type TEXT, and decode from type TEXT to type BYTEA.

  • **Added support for EXCLUDE columns in SELECT **

    Added support for EXCLUDE columns in SELECT * to define which columns to exclude from a SELECT wildcard expansion.

  • New setting for parsing literal strings

    New setting standard_conforming_strings controls whether strings are parsed without escaping, treating backslashes literally.

  • New keyboard shortcuts (UI release)

    Use new keyboard shortcuts in the SQL workspace to save and close scripts, and expand or collapse the results pane.

    • Close the current script in the SQL workspace with Ctrl + Alt + x for Windows & Linux, or ⌘ + Option + x for Mac
    • Close all scripts in the SQL workspace with Ctrl + Alt + g for Windows & Linux, or ⌘ + Option + g for Mac
    • Close all scripts except the one you are working on in the SQL workspace with Ctrl + Alt + o for Windows & Linux, or ⌘ + Option + o for Mac
    • Save the current script in the SQL workspace with Ctrl + Alt + s for Windows & Linux, or ⌘ + Option + s for Mac
    • Expand or collapse the results pane in the SQL workspace with Ctrl + Alt + e for Windows & Linux, or ⌘ + Option + e for Mac

Enhancements, changes, and new integrations

  • Added support for CREATE TABLE as an alias for CREATE FACT TABLE

    Added support for CREATE TABLE syntax, with the default as fact table. PRIMARY INDEX is now also optional for fact tables.

  • Added support for the DECIMAL data type with the ARRAY_SORT function

    ARRAY_SORT has been added as a function supporting the DECIMAL data type.

  • Minimize results in the SQL workspace (UI release)

    The results pane in the SQL workspace can now be minimized. Expand or collapse by double-clicking on the “Results” pane header, using the height control button to drag and change the size of the pane as desired, or using the keyboard shortcut Ctrl + Alt + e for Windows & Linux, or ⌘ + Option + e for Mac.

  • Primary index columns highlighted in columns object viewer (UI release)

    The columns pane in the object viewer now highlights columns that are part of a table’s primary index, making it easier to identify primary indexes and the order of the columns in the primary index.

Resolved issues

  • Fixed an issue where COPY TO export file size was limited to 2GB.

  • Fixed an issue that allowed more than one argument in CONCAT.

  • Returns an error when dynamic functions are used in aggregating indexes.

December 2022

New features

  • Added support for CSV TYPE options on ingest (DB version 3.14.0)

Added support for additional TYPE options for the CREATE EXTERNAL TABLE command, to allow configuration for ingesting different CSV file formats. Some of these options may be available in previous versions.

  • [ALLOW_DOUBLE_QUOTES = {TRUE|FALSE}], [ALLOW_SINGLE_QUOTES = {TRUE|FALSE}]: Define that unescaped double or single quotes in CSV input file will not cause an error to be generated on ingest.

  • [ALLOW_COLUMN_MISMATCH = {TRUE|FALSE}]: Defines that the number of delimited columns in a CSV input file can be fewer than the number of columns in the corresponding table.

  • [ALLOW_UNKNOWN_FIELDS = {TRUE|FALSE}]: Defines that the number of delimited columns in a CSV input file can be more than the number of columns in the corresponding table.

  • [ESCAPE_CHARACTER = {‘<character>’|NONE}: Defines which character is used to escape, to change interpretations from the original.

  • [FIELD_DELIMITER = '<field_delimeter>']: Defines a custom field delimiter to separate fields for ingest.

  • [NEW_LINE_CHARACTER = '<new_line_character>']: Defines a custom new line delimiter to separate entries for ingest.

  • [NULL_CHARACTER = '<null_character>']: Defines which character is interpreted as NULL.

  • [SKIP_BLANK_LINES {TRUE|FALSE}]: Defines that any blank lines encountered in the CSV input file will be skipped.

Enhancements, changes and new integrations

  • Added support for nullable arrays (DB version 3.11.0)

    Nullable type is now generally available for arrays, enabled for DB version 3.11 and above.

November 2022

New features

  • Added support for functions (DB version 3.13.0)

    • FIRST_VALUE: Returns the first value evaluated in the specified window frame.
    • NTH_VALUE: Returns the value evaluated of the nth row of the specified window frame (starting at the first row).
    • NTILE: Divides an ordered data set equally into a specified number of buckets.
    • CUME_DIST: Calculates the relative rank (cumulative distribution) of the current row in relation to other rows in the same partition within an ordered data set.
    • PERCENT_RANK: Calculates the relative rank of the current row within an ordered data set.
    • PERCENTILE_CONT (aggregation function): Calculates a percentile, assuming a continuous distribution of values.
    • PERCENTILE_CONT (window function): Calculates a percentile over a partition, assuming a continuous distribution of values.
    • PERCENTILE_DISC (aggregation function): Returns a percentile for an ordered data set, equal to a specific column value.
    • PERCENTILE_DISC (window function): Returns a percentile over a partition for an ordered data set, equal to a specific column value.

  • Added support for TRUNCATE TABLE command (DB version 3.11.0)

    Use the TRUNCATE TABLE command to remove all rows from a table.

  • Added support for DECIMAL data type (DB version 3.13.0)

    Beta support for the DECIMAL data type is coming in version 3.13.

    In previous versions, DECIMAL type columns are stored as DOUBLE type. Therefore, this change may require your action. Restart analytics engines before general purpose engines to use this new feature, and see below for additional actions.

    If your existing data model contains tables defined with DOUBLE type columns:

    • If you want to preserve the DOUBLE data type going forward, no change is required. The DECIMAL data type can be used for new tables/columns.

    If your existing data model contains tables defined with DECIMAL type columns:

    • If you do not want to change your tables, no action is required. These columns will just show up as type DOUBLE. You can safely use the DECIMAL type for new tables/columns.

    • If your table defined with columns as DECIMAL are recreated periodically (by your ELT process), new columns will be defined as DECIMAL(38,0) – default precision and scale. If you don’t want to switch to the DECIMAL data type for the existing data flows, you can change your ELT flows and replace the DECIMAL keyword with DOUBLE.

    • If you are using a function with a DECIMAL type in your ELT process, ensure that function is supported for the DECIMAL data type.

    • To change the data type of columns defined as DECIMAL but stored as DOUBLE, you will need to recreate the table with the new definition (column defined as DECIMAL(p,s)). To avoid precision loss, we highly recommend re-ingesting the data from the source (i.e., via external table) rather than casting values to decimal from the existing table (i.e., INSERT INTO new_table(d) SELECT CAST(d as DECIMAL(38,9)) FROM old_table;).

    • To avoid any downtime for the end-users in scenarios involving re-creating and re-ingesting the table, we suggest creating a view that reads from the existing table (old_table). After the new table (new_table) is created and data is ingested, repoint the view to the new table name (new_table) using the CREATE OR REPLACE VIEW command.

    Examples of changes in behavior

    • Comparing two values that differ by a small offset yields different results when stored as DECIMAL vs. DOUBLE, because DOUBLE is a variable-precision data type, whereas DECIMAL is exact with fixed precision: 
        SELECT '1.00000000000000000000000005'::DECIMAL(38,30) = '1.00000000000000000000000004999'::DECIMAL(38,30); -- false
  SELECT '1.00000000000000000000000005'::DOUBLE = '1.00000000000000000000000004999'::DOUBLE;  --true
    • Expressions involving DECIMAL data types with different precision and scale will yield an error: 
      SELECT 3::DECIMAL + 2::DECIMAL(3,1); -- Invalid operation error: Operations between decimals with different precision and scale is not supported at the moment. Explicitly cast one of the decimals to other decimal precision and scale.
    • Functions ROUND, TO_STRING, and TO_TEXT applied on DECIMAL will return the data type matching the input data type.

Enhancements, changes and new integrations

Resolved issues

  • Fixed an issue that caused degraded performance of the cache evictor. (DB version 3.13.0)

October 2022

New features

  • Added support for functions REGEXP_REPLACE and REGEXP_REPLACE_ALL (DB version 3.11.0)

    Use these functions to replace matching patterns in the input with a replacement. The REGEXP_REPLACE function replaces the first match only (from the left), REGEXP_REPLACE_ALL function replaces all the matches.

Enhancements, changes and new integrations

  • Added support for “OR” operator for JOIN (DB version 3.11.0)

    Allows performing JOINs with multiple join conditions linked via the “OR” operator

  • Updated syntax to generate an aggregating index (DB version 3.11.0)

    The CREATE AGGREGATING INDEX command will now generate the aggregating index, without using the additional AND GENERATE clause.

  • Added support for window function frame definitions (DB version 3.11.0)

    Add support for n PRECEDING, n FOLLOWING, and CURRENT ROW when defining the frame start and end in window functions.

  • VERSION() function now available (DB version 3.8.0)

    Query the engine version using the new VERSION() function. Engine version is also now available as a column in the information_schema.engines view.

  • Information schema updated (DB version 3.8.0)

    System-defined tables metadata can now be queried via the information_schema.tables view.

August 2022

Enhancements, changes and new integrations

  • Firebolt .NET SDK now available

    The Firebolt .NET SDK can be used to connect to Firebolt from .NET applications. The SDK supports .NET Core and is available at NuGet FireboltNetSDK.

  • Firebolt Go SDK now available

    The Firebolt Go SDK can be used to connect to Firebolt from Go applications. The SDK implements the Go database/sql driver specification. For more details, see the package listing at Firebolt Go SDK.

  • Information schema updated (DB version 3.7.0)

    query_history and running_queries views can now be queried via the information_schema.

    For more information, see Information schema for query history and Information schema for running queries.

  • Added support for Multi-factor authentication (MFA) (Beta)

    Firebolt now supports Multi-factor authentication.

    You can enable MFA for users or groups with the Firebolt Manager which sends a link to enroll in MFA using a QR code. When a user enrolls in MFA from the email, the status in Firebolt updates to MFA enabled.

    To enable MFA for a Firebolt user or group of users choose the User Management icon in the navigation pane. If the icon isn’t available, you don’t have Account Admin permissions.

    User management icon

    For more information, see Configuring MFA for users (Beta).

  • Added support for the hll_count_distinct(input, [, precision]) function (DB version 3.7.0)

    Allows for precision control of the count(distinct ) function with an optional precision parameter.

    Requires less memory than exact aggregation functions, like COUNT(DISTINCT), but also introduces statistical uncertainty. The default precision is 12, with a maximum of 20.

  • Added new data type aliases (DB version 3.7.0)

    Data type aliases have been added for REAL, FLOAT4, FLOAT8, INT4, INT8, and FLOAT(p). For more information on data types and their aliases (synonyms), see Data types.

  • Updated INFORMATION_SCHEMA.COLUMNS (DB version 3.8.0)

    Now includes more metadata on columns, as well as columns for views in a given database.

  • New script processing status in browser tab

    Added a status indicator in the browser tab so when multiple tabs are open in the browser, you can switch to a different tab and still track the status of your running script. The status adds a color coded indicator dot to the Firebolt icon in the tab. A green dot indicates the script completed successfully. The status remains in the tab for one minute after the script completes running.

    For more information about this new status indicator, and running scripts, see Running scripts and working with results.

  • Added dark mode

    Firebolt now supports an additional color theme - dark mode. You can toggle between light and dark modes in the UI. Select the toggle at the bottom of the left navigation pane to turn dark mode on and off.

  • Added support for IP allowed & blocked lists (Beta)

    Allows access to your Firebolt account from specific IP addresses. For more information, see Allowing and blocking source IP addresses for users (Beta)

  • Added support for server-side asynchronous querying on the Python SDK

    Allows the connection to be closed while the query is still running in Firebolt.

  • Released JDBC 2.0

    The Firebolt JBDC driver is now open-source under Apache 2. The new release also adds support for view tables in getTables().

Resolved issues

  • An error message is now displayed when too many partitions are added using a single INSERT statement.

July 2022

Enhancements, changes and new integrations

  • Added support for Airbyte

    The Firebolt integration includes an Airbyte Source connector as well as an Airbyte Destination connector, which allow you to easily move data into and out of Firebolt.

    For more information on Airbyte, see Airbyte.

  • New billing breakdown in engine dashboard

    The billing breakdown in the engine dashboard can now show billing or running time.

    Billing time

  • Added an optional <format> parameter to the TO_DATE and TO_TIMESTAMP functions

    The <format> parameter allows you to use a string literal, as shown in DATE_FORMAT, to specify the format of the string to convert. This hint helps the date-time parser to improve performance. For more information, see TO_DATE and TO_TIMESTAMP.

Resolved issues

  • Fixed an issue that prevented values from being set properly for a column with a DEFAULT constraint when an INSERT INTO statement ran without specifying the column value.

    For example, consider the following fact table and INSERT INTO statement.

    CREATE FACT TABLE t8
(
     col1 INT  NULL ,
     col2 INT  NOT NULL UNIQUE,
     col3 INT  NULL DEFAULT 1,
     col4 INT  NOT NULL DEFAULT 3,
     col5 TEXT
)
PRIMARY INDEX col2;

INSERT INTO t8 (col1,col2,col3,col5) VALUES (1,1,1,'a');

    Before the fix, SELECT * FROM t8 would return (1,1,1,0,'a'). The statement now returns the expected values, (1,1,1,3,'a').

  • Fixed a case where the APPROX_PERCENTILE function returns an error when running over an expression derived from multiple tables.

  • JSON_EXTRACT now supports the STRING data type as the expected type in addition to the TEXT data type.

  • Fixed an issue where a username that contained special alphabetical characters was not recognized when registering a Firebolt user.

  • Concatenating multiple arrays using the || operator now produces an array containing the input array values.

  • Fixed an issue where an array of LONGs could be displayed as an array of TEXTs in the SQL Workspace.

  • Fixed an issue that prevented a database without engines from being edited.

May 31, 2022

New features

  • You can now use the help menu to check the Firebolt service status page.
    Status Page

  • Added support for CREATE AND GENERATE AGGREGATING INDEX IF NOT EXISTS. For more information, see CREATE AGGREGATING INDEX.

  • Added an information schema view for indexes. The view is available for each database and contains one row for each index. For more information, see Information schema for indexes.

  • Added ARRAY_AGG as an alias of NEST. For more information, see NEST.

  • You can now concatenate strings, numbers, and arrays using the || operator without exlicitly casting elements.

  • Added TO_TEXT as an alias of TO_STRING. For more information, see TO_STRING.

  • An improved approach to Window functions is available for Beta testing by request. For more information, contact Firebolt Support.

Enhancements, changes and new integrations

  • Improved query execution on queries that read table data in Amazon S3. Firebolt now achieves greater read parallelization, which significantly speeds up queries that read many columns.

  • Cube.js integration
    You can now configure a Firebolt database as a data source for Cube.js. Cube.js is an embedded BI platform that allows you to define a data model, manage security and multitenancy, accelerate queries, and expose data to your applications using SQL and APIs. For more information, see Firebolt in Cube.js documentation.

  • Tableau connector
    An official Firebolt connector is now available in the Tableau Exchange so you can use Tableau more easily to analyze and visualize your Firebolt data.

  • Sifflet integration
    Firebolt now integrates with Sifflet to improve your visibility into data quality issues, data lineage, and data discovery. For more information, see Firebolt in Sifflet documentation.

  • Looker symmetric aggregates
    Symmetric aggregates in Looker are now supported when connected to Firebolt. For more information, see A Simple Explanation of Symmetric Aggregates in the Looker Help Center.

  • dbt append-only incremental materializations
    The dbt-firebolt adapter now supports append-only incremental materializations. For more information, see Firebolt configurations in dbt documentation.

Resolved issues

  • Fixed an issue that caused a Type mismatch in joinGet error when using a join index in table joins where columns are of different numeric types—for example, joining a column in a fact table defined as DOUBLE with a column defined as BIGINT in a dimension table.

  • Fixed an issue that caused the ARRAY_JOIN function to fail with an array of numeric values.

  • Fixed an issue in the REGEXP_MATCHES function that caused an error in some circumstances when a specified pattern didn’t match the input string.

  • Fixed an issue that caused a session to hang when trying to query an external table referencing an empty JSON file in Amazon S3.

  • Fixed an issue that caused incorrect column names when using COPY TO with a SELECT statement that uses aliases.

May 11, 2022

New features

  • Added support for CREATE OR REPLACE VIEW statement
    The IF NOT EXISTS and OR REPLACE options are incompatible when creating a view. For more information, see CREATE VIEW.

  • Added ALTER DATABASE statement
    Allows you to change the engine configuration and the database description. For more information, see ALTER DATABASE.

  • Added TO_CHAR function
    Converts a timestamp or a numeric type to strings in a format that you specify. For more information, see TO_CHAR.

  • Materialized CTEs (Beta)
    Added an optimization hint to support caching large, computationally expensive CTE results in memory for reuse to increase query performance. For more information, see SELECT.

  • Added support for SET statements passed via SQL in the Python SDK

Enhancements, changes and new integrations

  • Engine specs are now Firebolt-specific
    Firebolt engine spec names have changed from AWS EC2 names to Firebolt-specific names—for example, M4. This change simplifies engine specs and aligns them with Firebolt usage and deployment. This is the first step toward improving the overall engine selection and usage experience, allowing Firebolt to update underlying hardware more quickly and automatically. For more information, see Working with engines.

  • Expanded application of aggregating indexes
    The query optimizer now uses aggregating indexes in more cases to improve performance.

  • Python SDK updates
    • Support added for DECIMAL, DATE32, and DATETIME64 data types in the Python SDK.
    • Added the ability to disable caching of authentication tokens for use in read-only environments like AWS Lambda.
  • Added a materialization connector in Estuary for Firebolt
    For more information, see the Firebolt materialization connector topic in Estuary documentation.

Resolved issues

  • Fixed an issue where complex join conditions could fail when evaluating NULL values.

  • Fixed an issue that caused dropping partitions by date extraction to fail.

  • Fixed an issue that caused DROP AGGREGATING INDEX to run slower than expected.

  • Fixed an issue with queries containing subqueries, where the outer query limit clause would override the inner query limit clause.

  • Fixed an issue that could cause canceling a query to fail if the query was reading an external table.

  • Fixed an issue that could prevent aggregating index creation on aggregations used with a few functions (for example, SUM, LENGTH, and COUNT).

  • Fixed an issue that caused NULLs to be mishandled in some CASE statements.

  • Fixed an issue that required non-reserved keywords to be enclosed in quotes.

  • Fixed the data type for the number_of_rows column in information_schema.tables from INT to BIGINT.

March 30, 2022

Improved error message experience

In the SQL workspace, you can now select a portion of an error message to copy.

Select partial error message

March 15, 2022

New features

  • Visual explain
    Use the new visual explain feature to analyze query execution plans. Different display formats and navigation options enhance your ability to quickly analyze and diagnose complex queries. For more information, see Analyze query execution plans with visual explain.

  • COPY TO statement (Beta)
    Use the new COPY TO (Beta) statement to export query results to an Amazon S3 location in CSV, TSV, JSON, or Parquet file formats. Configuration options allow you to choose file output options. For more information, see COPY TO (Beta).

  • Firebolt CLI now available
    Run SQL and manage databases and engines in Firebolt directly from the command line. Use interactive mode or invoke commands from scripts. For more information, see firebolt-cli on PyPI.

  • Python SDK now supports parameterized queries and multiple statements
    • Parameterized queries – Allows parameters to be inserted dynamically into SQL. Handles sanitization to help prevent SQL injection attacks. For more information, see .execute() in the Python Database API Specification and firebolt.db.cursor module in the Firebolt Python SDK reference.
    • Multiple statements – Parses and separates multi-statement SQL, and then runs statements sequentially. For more information, see .executemany() and .nextset() in the Python Database API Specification, and firebolt.db.cursor module in the Firebolt Python SDK reference.

  • Node.js SDK now available
    Provides a database driver for Node.js applications to connect to Firebolt and run queries more easily. For more information, see Firebolt Node.js SDK on npm.

  • Information schema for views
    For more information, see Information schema for views.

  • Information schema for engines
    For more information, see Information schema for engines.

  • Added ATAN2(<y_expr>,<x_expr>)
    Added the two-argument arc tangent function ATAN2.

Enhancements, changes, and resolved issues

  • Breaking change
    • Floating point data type columns are no longer supported in partition key definitions.

  • Enhanced Parquet support
    Firebolt now supports Parquet list fields with a period (.) in the name. In addition, Parquet nested lists are now supported.

  • catalog.query_history limited to 14 days
    Queries older than 14 days are no longer available when querying catalog.query_history.

  • Firebolt dbt adapter
    • Improved connection method – The Firebolt dbt adapter now connects to Firebolt using the Python SDK instead of JDBC. JDBC and the JRE are no longer required.
    • Upgraded to dbt 1.0.0 – The Firebolt dbt adapter has been upgraded to v1.0.0. For more information about new features available in 1.0.0, see New features and changed documentation in dbt documentation.
  • Resolved issues
    • Fixed an issue that caused inaccurate elapsed time for queries to be shown in the SQL workspace.
    • Fixed an issue during INSERT INTO operations that caused some engines to crash when reaching out of memory instead of failing gracefully.
    • Fixed an issue that caused the cached data ratio to be inaccurate for queries with a high LIMIT value.
    • Fixed an issue that could cause a DROP AGGREGATING INDEX statement to hang.
    • Fixed an issue that caused an engine crash when an unsupported function was used in a partition key definition.
    • Fixed an issue that allowed an aggregating index to be created with a window function, which is unsupported and caused unpredictable issues. Creating an aggregating index with a window function now fails with an error.

February 16, 2022

  • Fixed an issue where view creation could fail when the view depended on other views.

January 31, 2022

What’s new

UI

A new help button allows you to contact Support directly through a support form and access release notes and documentation on the web.

The Contact Support form pre-populates your contact information (Name, Company, and E-mail). It allows you to select the issue severity, provide details of the affected engine, and provide details about the issue or question.

Jan 20, 2022

What’s new

SQL and warehouse

Added the following functions to handle NULL values:

  • IFNULL ( <expr1>, <expr2> ). If <expr1> is NULL, returns <expr2>; otherwise, returns <expr1>.

  • ZEROIFNULL ( <expr> ). If <expr> is NULL, returns 0; otherwise, returns <expr>.

Bug fixes

  • Fixed an issue that could cause a failure when canceling a query run on external tables.

  • Fixed an issue that prevented using the REPLACE keyword within the CREATE VIEW statement.

Dec 16th, 2021

What’s new

SQL and database

  • New BTRIM, RTRIM, and LTRIM functions support removing a specified string from both sides (BTRIM) the right side (RTRIM) or left side (LTRIM) of a source string.

    Each of these functions accepts arguments in the form (<source_string>[, <chars_to_trim>]). If <chars_to_trim> is omitted, a space character is assumed.

    For example, each of the following clauses returns the string ‘trim’.
    BTRIM('xyxtrimyyx', 'xy')
    LTRIM('zzzytrim', 'xyz')
    RTRIM('trimxxxx', 'x')

  • Query execution has been improved through faster query parsing and caching of query plans.

  • max_execution_time is now available as a set parameter for controlling a timeout threshold for SQL statements.

UI

You can now use a keyboard shortcut to create a new script in the SQL workspace (Windows – Ctrl+Alt+N, Mac – Cmd-Opt-N).

Nov 23rd, 2021

What’s new

UI

  • Added an indication of the active script in the browser tab and SQL workspace.

  • Improved resizing behavior of the database panel in the SQL workspace.

  • Added an option that allows you to filter query results by numeric and text values that you provide in addition to possible column values.

  • Added a button that allows you to turn background filtering on or off. By default, background filtering is on—filtered results are affected by other filters. For example, if you filter by city “Stockholm”, then by customer name “Jones”, you see only query results for Joneses in Stockholm. You can now use this button to toggle background filters on and off.

Nov 22nd, 2021

What’s new

SQL & warehouse

  • Firebolt now optimizes semi-joins against dimension tables when row uniqueness is identified.

  • Repeated SQL statements now reuse query plans, which improves parsing performance.

  • You can now use GEN_RANDOM_UUID to generate a version 4 universally unique identifier (UUID) as a STRING data type.

  • For engines that are always running, the catalog.query_history view now has a TTL of 14 days.

Integrations and connectors

  • Released the new Python SDK, which implements the Python Database API v2.0 specification on top of the Firebolt REST API. The SDK makes it easier to connect and run SQL queries on Firebolt from a Python application.

  • Released the new SQLAlchemy adapter for Firebolt, which allows applications to connect and run SQL queries on Firebolt through the SQLAlchemy interface.

  • Released the new dbt adapter for Firebolt, which allows data analysts and data engineers to build and operationalize data transformations in Firebolt.

  • Added integrations for Superset, Keboola, Hevo, and Nexla.

Bug fixes

  • Fixed an issue that prevented views without a defined schema from being used in table creation via CTAS.

  • Fixed an issue where queries could fail to run due to insufficient resources.

  • Fixed an issue where engines with the warmup method set to preload all data could fail.

Nov 10th, 2021

What’s new

  • Added the REGEXP_MATCHES function, which returns an array of all of the captured substrings resulting from matching a given regular expression pattern.

  • Added the TO_WEEKISO function that returns the week according to ISO rules.

  • Added support for theta joins using interval data types.

  • Improved query compile time. In some cases queries will compile 50% faster or more.

  • Added natural join support.

Bug fixes

  • Fixed an issue with APPROX_PERCENTILE where multiple uses within the same query could fail.

  • Fixed an issue where a long list of values used with INSERT could fail.

  • Fixed an issue where using S3 metadata columns in a subquery could fail.

Sep 2nd, 2021

What’s new

  • You can now see the percentage of warm data per query in the SQL workspace.

  • You can now search for a table or column by its name in the database panel within the SQL workspace.

Bug fixes

  • Fixed an issue where canceling an engine startup process required an initial warmup to complete.

  • Fixed an issue where general-purpose engines with the Preload all data warmup method could fail to start.

Aug 12th, 2021

What’s new

  • You can now quickly investigate statistics for query results in the SQL Workspace. Statistics such as the number of unique values, number of empty values, min/max/median and other quartiles, data ranges, top/bottom numbers, and more are available. Click the expansion arrow to the left of the header row to see statistics. For more information, see Viewing query statistics with results in Firebolt documentation.

  • You can now sort and filter data values on any column in your query result-set in the SQL Workspace.

  • The Firebolt Manager now displays when a database is being dropped.

  • The Engines page has been discontinued. You can create and manage engines when you edit a database, and engines not has_toc: false attached to a database are no longer supported.

Bug fixes

  • Fixed an issue where using the JSON_EXTRACT function to extract arrays could fail.

  • Fixed an issue where engines could take a long time to start due to an infrastructure failure.

  • Fixed an issue where engines could fail to start due to corrupted views.

July 29th, 2021

What’s new

  • Firebolt now supports statements that contain INSERT INTO with GROUP BY on an engine with multiple nodes.

  • The third parameter in SUBSTR is now optional. When omitted, the function returns the entire string to the right of the offset.

Bug fixes

  • Fixed an issue where using the SPLIT_PART function without specifying all the mandatory parameters could cause an engine crash.

  • Fixed an issue where returned rows & returned bytes columns in the catalog.query_history view could mistakenly contain zeros.

  • Fixed an issue where canceled queries could be marked as failed in the catalog.query_history view.

  • Fixed an issue where CTAS could fail to infer nullability correctly.

June 24th, 2021

What’s new

  • You can now upload a script file (*.sql or *.txt) from your computer to the SQL workspace

  • Formatting scripts in the SQL workspace is now supported (Mac: CMD + OPT + L, PC: CTRL + ALT + L)

  • Account admins can now perform bulk operations on users, making it easier to:

    • Remove user(s)
    • Accept approval
    • Deny approval
    • Resend email
    • Cancel invitation

May 27th, 2021

What’s new

  • The boolean data type is now supported

  • Extracting quarter, day of the week, day of the year, hour, and the week is now possible when using the EXTRACT function

Bug fixes

  • In some cases, % at the beginning of the like value was ignored when an escape character was used.

  • In some cases, ingested data took longer to propagate to analytic engines.

May 5th, 2021

What’s new

SQL & Warehouse enhancements

  • Firebolt engines have a new failover mechanism for nodes. Engines can repair failed cluster nodes automatically for increased reliability

  • Configuring the warmup method is now supported when creating an engine using SQL.

  • The CURRENT_DATE keyword is now supported.

  • Using identical aliases across different sub-queries is now supported.

  • Casting using the :: notation is now supported.

UI/UX enhancements

  • Easily reorder scripts in the SQL Workspace by dragging the script tab to the desired position

  • Download SQL scripts from the SQL Workspace to your computer

  • You can now clear all filters at once in the Databases page

  • The Account & Billing page was updated, and it now contains a cost breakdown for compute and storage in addition to the AWS account ID

  • Updated the design of the results table in the SQL Workspace

  • Deleting a database now deletes the engines attached to it

  • The SQL Workspace now lists corrupted tables in orange with a warning icon

Bug fixes

  • Opening the engines drop-down menu when a script was running was not possible

  • Numeric query results could have been mistakenly rounded in the UI

  • Querying external tables with predicates on the virtual columns could fail

  • Running the SHOW CACHE command could fail

  • Using AWS role-based authentication in external tables could fail

  • The ‘Inserted rows’ value in the query history table could be higher than expected

  • Creating an engine using SQL could result in a general-purpose engine although a data analytics engine was requested

Apr 22nd, 2021

What’s new

  • Configuring the warmup method is now supported when creating an engine using SQL

  • A new JDBC driver is now available with support for canceling a running query

  • New window functions: LAG, LEAD, AVERAGE, ROW_NUMBER.

  • Aliasing sub-queries is now optional

  • Skipping header rows in files when loading data into Firebolt is now supported

  • Implicit cast from text column with textual numeric values into numeric data types was added

Bug fixes

  • Loading data into Firebolt tables could fail due to background network issues under some circumstances

  • Running COUNT(*) and COUNT(1) on the same data set could result in different results

  • An engine could crash when issuing a drop table command after a failed ingestion

  • Using an aggregated value as a single predicate within the WHERE clause could fail

  • Using the COUNT window function with the NEST function could fail

  • Data loss could happen in some rare cases when loading data in async mode following by an engine restart

Mar 25th, 2021

What’s new

  • Create table as select (CTAS) is now available

  • Added information schema for databases, tables, and columns

  • AWS Glue can now be used with external tables to simplify schema creation

Bug fixes

  • Loading data into Firebolt tables could fail due to network issues in some rare circumstances

  • Aggregating indexes were not being used for queries containing a WITH clause

  • Running SHOW INDEXES command contained N/A under the index size column for join indexes

  • Creating a new database and removing the default engine could fail

  • WHERE statements could sometimes be ignored when using views or CTEs (Common Table Expressions)

Feb 25th, 2021

What’s new

New databases page

We’ve updated the databases page for a smoother experience:

  • Easily find the databases and engines that you need with new elaborate search and sort options

  • It is now possible to view the page in two modes: detailed and summarized

  • Performing operations on engines quickly - start, stop or edit an engine directly from the database

  • Quickly jump to the SQL workspace from a selected database

Additional UI goodies

  • We added a query history script to the list of quick-start templates in the SQL Workspace

  • The script results section has been enhanced to show the full query text with richer syntax highlighting, expand/collapse modes, and indentation

New SQL functions

  • MD5_NUMBER_LOWER64
  • MD5_NUMBER_UPPER64
  • REPEAT

Feb 11th, 2021

What’s new

New functions

  • New window functions are now supported: MIN, MAX, RANK, DENSE_RANK

  • The following DATE functions are now available:
    • TO_DAY_OF_WEEK
    • TO_DAY_OF_YEAR
    • TO_YEAR
    • TO_QUARTER
    • TO_MONTH
    • TO_WEEK
    • TO_HOUR
    • TO_MINUTE
    • TO_SECOND

  • Calculating the MD5 hash of a string is now supported using the MD5 function

  • The engine tooltip in the database editor was enhanced to include its specifications such as ec2 type, number of nodes, warmup method, etc

Bug fixes

  • Join index creation could fail with an out-of-memory- error due to over allocation of RAM

  • SELECT DISTINCT queries could fail

  • When using SHOW TABLES the primary index column would show wrong results when an aggregating index was in place

  • COUNT DISTINCT with INNER JOIN could return incorrect results

  • Using the TO_TIMESTAMP function could fail when parsed with certain date formats

Jan 26th, 2021

What’s new

New database & engine creation experience

We’ve updated the way engines and databases are created and edited, for a smoother experience:

  • Engines and databases are now created and managed together in one fluent experience

  • Engines can now only be created as part of a database

  • When creating a new database, you can quickly add multiple engines to it

  • Engine scale now supports any number between 1-128, not just fixed multiples of 2

  • The EC2 type selection has been improved with a clearer categorization of EC2 family types

Jan 21st, 2021

What’s new

  • OFFSET clause is now supported

  • New optimization for joins - When filtering on a dimension table in a fact-to-dim table join queries are now much faster thanks to optimized multi-predicate pushdowns

  • SHOW CACHE now shows you detailed information on the engine’s SSD utilization

  • Our JDBC driver has been updated with multiple bug fixes and improvements

Bug fixes

  • A failed generate aggregating index command could be mistakenly reported as successful

  • Aggregating index creation could fail due to ‘file name too long’ error

Dec 31st, 2020

What’s new

New functions:

  • INTERVAL is now supported for arithmetic operations on dates

  • The following JSON functions were added: JSON_EXTRACT, JSON_EXTRACT_KEYS, and JSON_EXTRACT_VALUES

  • The following array functions were added: ARRAY_MIN, ARRAY_MAX

Performance and Engines:

  • The new AWS EC2 Graviton2 nodes (c6gd/r6gd/m6gd families) are now available for use in your engines

  • Join indexes now boost your query performance even further

  • GROUP BY queries have been improved to run faster through enhanced aggregate-pushdown utilization

UX enhancements:

  • New look and feel for the application toolbar, introducing a lightweight skin that enables more work space with clear visibility

  • Introducing a new database selection and script management panel in the SQL Workspace next to the application toolbar, including collapsed/expanded modes

    • The SQL Workspace database selection and script management panel can now be resized

Bug fixes

  • An aggregating index could be accidentally dropped when recreating the same index sequentially

  • Queries using the WHERE IN clause could run slower than expected

Dec 16th, 2020

What’s new

  • Engine selection is now available when using our JDBC driver

  • Ingesting TSV files is now supported

  • Editing the engine’s warmup method is now supported using SQL

  • New search boxes have been added to the engine and database pages for quick access

  • Application state is now preserved between browser sessions, keeping the last page and client settings such as the SQL, active SQL tab, sort order, etc

Bug fixes

  • Using the INSERT INTO VALUES statement could fail

  • Querying using the LIMIT clause could fail

  • Long queries having query plan weighing more than 1MB could fail

  • SQL error with long text can’t be scrolled to view the entire error description

Dec 15th, 2020

The wonderful point in time in which a new product decides to start sharing release notes.