0.33.0

[0.33.0] - 2026-04-20

Added

• Column hints for the update! / u! operation. Typing u! or u! col = val, now suggests remaining assignable columns, excluding those already assigned.

Changed

• The => count in the group operation is now optional. count is used by default when omitted:

email | g: status

[0.32.0] - 2026-03-30

Added

• Multi-table update! support: when assignments target different tables (e.g. c.deleted_at and d.deleted_at), multiple UPDATE queries are run—one per table.
• API eval response for update! now includes per-table results: [["Table" "Rows updated"] ["company" 5] ["document" 3]].

Fixed

• Recursive delete no longer follows heuristic relations — only real foreign key constraints are traversed. Heuristic relations are now flagged in ast.hints.table via a heuristic boolean so clients can distinguish them.
• update! now uses the table alias when columns are qualified (e.g. c.name), so updates target the correct table when multiple tables are in context:

company as c | w: id = 1 | document | w: type = 'invoice' | update! c.deleted_at = '2026-01-01'

[0.31.0] - 2026-02-16

Added

Build endpoint returns:

• Prettified expression in the ast.prettified property.
• Ranges for the operations in the ast.ranges property.

[0.30.0] - 2026-02-04

Added

• Support for heuristic relations based on column naming conventions. This is helpful when foreign keys are not explicitly specified.

[0.29.0] - 2025-12-25

Added

• Support for cursor position aware hints. This is helpful when the user isn't at the end of the expression. Hints are generated based on the cursor position. The build endpoint supports a new parameter cursor which must contain the line and character position of the cursor.

[0.28.0] - 2025-12-08

Added

• Support for date extraction functions in the select operation e.g.

employee | select: created_at => year
employee | select: created_at => year as created_at_year

Supported functions are: year, month, week, day, hour, minute

• Group on derived columns e.g.

employee | select: created_at => month | group: month => count

• Explicit join columns are supported in the join operation e.g.

company | employee .company_id = .id

Removed

• Internal state field :join-map has been removed. This was legacy dead code kept since v0.8.0 that was never actually used. The :joins vector format continues to be used for SQL generation.

[0.27.0] - 2025-10-19

Added

• Column aliases are supported in the order operation e.g.

company as c | o: c.name asc

• Support comments in the expressions e.g.

company | -- This is a line comment
company | /* This is a multi-line block comment */

[0.26.1] - 2025-09-07

Changed

• Using a readonly db user for the playground

[0.26.0] - 2025-09-06

Added

• Support for raw SQL queries:

POST /api/v1/sql

[0.25.0] - 2025-08-28

Added

• Values are type casted to the appropriate database column type.

Fixed

• As we use the correct typecase for the values, it is possible to update a jsonb column.

• It is possible to use a LIKE operator on a uuid using a type cast e.g.

company | where: id like '9cd%' ::uuid

[0.24.0] - 2025-08-25

Added

• Support for update! operation:

customers | w: id = 1 | update! name = 'John Doe'

• Return id columns for all tables in the result. This allows in-place updates on a query result.

Fixed

• All columns were being returned in some cases instead of the explicitly selected columns. e.g.

company | s: id,
company | s: id | l: 1

[0.23.0] - 2025-08-15

Added

• Setup for playground

Fixed

• Numbers are parsed as longs e.g. if the column is an integer:

customers | id = 1

0.22.0

[0.22.0] - 2025-07-12

Added

• Column hints for where: operation, supporting partial expressions:

company | where:           # Shows all columns
company | w: i             # Shows columns matching 'i' (like 'id')  
company | w: id =          # Shows all columns after specifying column + operator
y.employee | w: comp       # Shows columns matching 'comp' (like 'company_id')

[0.21.0] - 2025-07-02

Fixed

• Docker image wasn't running.Updated the base image to openjdk:11-jre-slim

Added

• Support for ilike, not like, and not ilike operators:

company | where: name ilike 'acme%'
company | where: name not like 'test%'
company | where: name not ilike 'admin%'

• Support for casting columns as ::uuid
• Support for dates in conditions e.g.

company | where: created_at > '2025-01-01' | created_at < '2026-01-01'

0.20.0

[0.20.0] - 2025-06-22

Added

• Specify join types i.e. LEFT JOIN or RIGHT JOIN:

x | y :left
x | y :right

Breaking

• Syntax for specifying parent and child relations is changed (introduced in 0.6.0). This avoids the need for backtracking.

x | of: y
x | has: y

is now:

x | y :parent
x | y :child

• ^ is removed from the syntax to specific the directionality of the join. (introduced in 0.6.0)

[0.19.0] - 2025-06-21

Added

• Support for casting columns in conditions e.g.

company | where: id like '9cd%' ::text

[0.18.0] - 2025-06-04

Added

• Support for group operation:

email | group: status => count

• Column aliases are supported in conditions e.g.

tenant as t | company | where: t.id = 'xxx'

Changed

• Default limit is removed for count: and delete: operations.
• For count: operations, the with SQL clause is used to build the nested query e.g.

company | count:

is evaluated to:

WITH x AS (SELECT * FROM "public"."company") SELECT COUNT(*) FROM x;

0.17.0

[0.17.0] - 2025-05-03

Fixed

• Using database connection pooling
• Using UTC dates

0.16.0

[0.16.0] - 2025-02-09

Added

• Support for connection stats which contain the number of db connections.

GET /connection/stats

{
  "connection-count": 10,
  "time": "2025-02-10T01:49:53.808120858"
}

0.14.0

[0.14.0] - 2025-01-07

Added

• Support for columns e.g. hints are generated for a partial select:

company | s:
company | s: id,

Changed

• Connection id format is host:port instead of just the host.

[0.13.0] - 2024-10-25

Added

• DB connection management i.e. create a new connection and connect to it

POST /connections
POST /connections/:id/connect

• Support for booleans:

company | is_public = true

[0.12.0] - 2024-10-18

Added

• Support for not in operator
• Support for no operations e.g. delete:. Such operations are evaluated client side.

[0.11.0] - 2024-09-12

Added

• where: supports comparing values between columns of different tables

folder as f | document | where: name = f.name
folder as f | document | name != f.name

[0.10.0] - 2024-09-12

Added

• Support for count::

company | count:

0.9.0

Added

• Support for NULL:

company | name is null
company | name = null
company | name is not null

• Support for order:

company | order: created_at
company | order: country, created_at asc

0.8.1

[0.8.1] - 2024-07-30

Fixed

• Specifying the join column in case of ambigious relations wasn't working.

[0.8.0] - 2024-07-30

Added

• Change the context using the from: keyword. This is helpful when the tables relations are not linear and look like a tree.

company as c | document | from: c | employee

Breaking

• State: joins is a vector e.g. [ "x" "y" ["x" "id" :has "y" "x_id"]]

Changed

• State: join-map is kept for legacy reasons but it is only used internally.

[0.7.2] - 2024-07-26

Changed

• No difference in functionality. Removed a lot of deprecated code - only keeping the code for reborn.

[0.7.1] - 2024-07-26

Fixed

• Allow spaces in the start of a pine expression

[0.7.0] - 2024-07-26

Added

• Support for in operator

Changed

• Error type is returned. It is either nothing or parse.

[0.6.0] - 2024-07-22

Added

• Support for directional joins:

employee | has: employee
employee | of: employee
employee | employee^

• Columns can be qualified by table aliases:

employee as e | s: e.name

[0.5.4] - 2024-07-16

Fixed

• Incorrect hints were generated in case of ambiguity

[0.5.3] - 2024-07-16

Fixed

• Incorrect schema being returned in hints when joining from child to parent

[0.5.2] - 2024-07-14

Changed

• Default limit is 250 if not specified

Fixed

• All columns weren't being select in some cases e.g. using company | s: id | employee, the columns from employee table weren't being selected

[0.5.1] - 2024-07-11

Added

• Context sensitive columns selection e.g. company | s: id | employee | s: id

[0.5.0] - 2024-07-10

Added

• Hints can be provided to resolve ambigious joins e.g. instead of company | employee, you can explicitly specify the join column i.e. column | employee .company_id
• The delete operation uses a nested query. The column used for deletes must be specified:

public.company | delete! .id

evaluates to:

DELETE FROM
  "public"."company"
WHERE
  "id" IN (
    SELECT
      "c_0"."id"
    FROM
      "public"."company" AS "c_0"
  );

• Conditions can be composed. Following are allowed:

company | where: id='xxx'
company | w: id='xxx'
company | id='xxx'

Changed

• Conditions can't be combined with the tables e.g. company id='xxx'. Instead compose them using pipes: company | id='xxx'
• Double quotes around strings aren't supported anymore. Use single quotes i.e. instead of id="xxx", use id='xxx'

Removed

• Support for group, order, set! is dropped. It will be added soon in the up coming versions.
• Context sensitive columns selection

0.4.0

Added

• Disabled CORS
• API endpoint for getting the active connection:

GET /connection

{
  ...
  "connection-id": "..."
}

• When using POST /build the response also includes the connection-id, and params
• When using POST /eval the response also includes the connection-id, query, and params.
• In case of an error, it is handled and the error message is returned in the API response as error
• Limited support for showing hints based on the input

Fixed

• Pine expression build/eval was failing if the db connection isn't initialized
• An error was being thrown when using uuid values in the expressions: operator does not exist: uuid = character varying
• Order of the columns in the result was sometimes not the same as the order in the query. Also, all columns are explicitly selected in the sql instead of relying on *

0.3.1

What's Changed

• Feat/support multiple connections/1 by @ahmadnazir in #9
• Added endpoint for evaluating pine expressions by @ahmadnazir in #12
• Using json objects for responses by @ahmadnazir in #13
• Feat/set connection via api/1 by @ahmadnazir in #14
• API endpoint for getting connection ids to choose from by @ahmadnazir in #16

Full Changelog: 0.3.0...0.3.1